Small Satellites for Earth Observation: Selected Proceedings of the 5th International Symposium of the International Academy of Astronautics, Berlin, April 4-8 2005 9783110919806, 9783110188516

Citation preview

Small Satellites for Earth Observation

Small Satellites for Earth Observation Selected Proceedings of the 5 th International Symposium of the International Academy of Astronautics Berlin, April 4 - 8 , 2005

edited by Hans-Peter Röser Rainer Sandau Arnoldo Valenzuela

W DE G_ Walter de Gruyter · Berlin · New York

Editors Hans-Peter Röser Universität Stuttgart Institut für Raumfahrtsysteme Pfaffenwaldring 31 70550 Stuttgart Germany

Rainer Sandau DLR Rutherfordstr. 2 12489 Berlin Germany

Arnoldo Valenzuela MEDIA LARIO S.R.L. Localitä Pascolo 23842 Bosisio Parini (LC) Italy

® Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

ISBN-13: 978-3-11-018851-6 ISBN-10: 3-11-018851-1 Bibliographic information published by Die Deutsche

Bibliothek

Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the Internet at .

© Copyright 2005 by Walter de Gruyter G m b H & Co. KG, 10785 Berlin, Germany. All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Printed in Germany. Cover design: Martin Zech, Bremen.

Table of Contents Preface

Session 2 - Earth Observation Missions SMOS: An L-Band Interferometric Radiometer Mission Bermudo, F., Venet, M., Le Du, Μ., Kerr, Y., Barre, Η.

7

Earth Observation - a new paradigm Mostert, S„ du Plessis, J., Cronje, T.

15

BIRD Fire Recognition and Comparison with Terra/MODIS Oertel, D. , Lorenz, Ε., Zhukov, Β., Csiszar, I.

23

Earth Observation Results of FORMOSAT-2 from June 2004 to February 2005 Chern, J.-S.

31

Real Time Weather Forecasting: The EGPM Mission Concept by Carlo Gavazzi Space Team Lo Rizzo, V,, Ortix, F., Jarrett, M„ Erickson, P. D„ Warne, D. H.

40

Session 3 - Constellations & Platforms Progress in small satellite technology for earth observation missions da Silva Curiel, Α., Cawthorne, Α., Sweeting, M.

50

Technical aspects and attitude control strategy of LAPAN-TUBSAT micro satellite Hardhienata, S., Nuryanto, A„ Triharjanto, R. H., Renner, U.

64

OHB platforms for constellation satellites Koebel, D., Lübberstedt, Η., Tobehn, C., Penne, Β.

71

A multiple satellite concept for severe weather monitoring Drescher, Α., Sandau, R.

75

Session 4 - Results & Lessons Learned Star-tracking in high radiation regime: In-flight results from the SMART-1 mission Denver, T., Jorgensen, J. L„ Jorgensen, P. S., Rathsman, P.

81

In-orbit performance verification of FORMOSAT-2 Chern. J.-S.

88

BIRD: More than 3 years experience in orbit Lorenz. E„ Jahn, Η.

102

Demonstration of a Semi-operational Fire Recognition Service Using BIRD Microsatellite Lorenz, E„ Casanova, J. L„ Gonzalo, J., Oertel, D., Aguirre, M„ Leibrandt, W., Billing, G., Martin de Mercado, G.

110

Session 5 - Educational Programs Results of space weather research on scientific-educational micro-satellite "KOLIBRI-2000" Klimov. S. I., Grigoryan, O. R., Grushin, V. A... Novikov. D. I.. Petrov, V. L.. Savin. S. P.

119

BEOSAT (Brunswick-Earth-Observation-Satellite) Kluge, R., Pfingstgräff, J., Berger, Μ., Kubus, D. Proposal for a new hyper spectral imaging micro satellite: SVALBIRD Sigernes, F., Renner, U.. Roemer. S., Bleif, J.-H.. Lorentzen, D. Α.. Claes, S., Nordheim, R.. Johannessen, F. , Ekstrand, V., Pedersen, S.-C.

134

X-SAT Mission progress Bretschneider, T„ Tan, S. H„ Göll, C. H„ Arichandran, K„ Koh, W. E„ Gill, E.

145

Session 6 - Student Conference Payload and scientific investigation of the Flying Laptop Walz, S. , Lengowski, M„ von Schoenermark, M.

153

Remote satellite position & pose estimation using monocular vision Malan, D. F., Steyn, W. H„ Herbst, Β. Μ.

161

Remote sensing by University of Tokyo's pico-satellite project " P R I S M " Enokuchi, Α., Nagai, M., Funase, R., Nakamura, Y., Nakasuka, S.

169

Session 7 - Instruments Advanced instruments and their impact on earth science missions (1) Hartley, J. , Komar, G., Lemmerman, L., Gerber, A.

177

Advanced instruments and their impact on earth science missions (11) Gerber, Α., Lemmerman, L.. Hartley, J.

185

JSS multispectral imagers for earth observation missions Döngi, F.. Engel, W . . Pillukat, Α., Kirschstein. Ο.

192

Development of E a r t h Observation Sensors for Small Satellites in Satrec Initiative Kim, E.-E.. Choi, Y.-W.. Kang, M.-S., Jeong, S.-K.

200

The evaluation of un-cooled detectors for low-cost thermal-IR earth observation at the Surrey Space Centre Oelrich, B., Underwood, C„ Mackin, S.

206

Session 8 - Special Aspects Gun launch system for small satellites: a fresh look Degtyarev. Α., Kanevsky, V., Ventskovsky, O.

214

Development of System Engineering Design Tool (SEDT) for small satellite conceptual design Chang, Y.-K., Hwang, K.-L., Kang, S.-J., Moon, B.-Y., Kim, S.-J., Chang, J.-S.

222

The R O C K O T launch vehicle - f u t u r e missions and plans for enhancing cost-effective access to orbit for the small satellite community Kinnersley, M., Freeborn, P., Schumacher, I., Zorina, A.

230

Market analysis and demand forecasting methodology in project management of small satellites in South East Asia (SEA) Saeedipour, H. R„ Said, Md. A. Md., Allaudin, M. F.

238

Aircraft-Based Satellite Launching (ABSL) System - Future Space Transportation System Kamal Adnan, M. S„ Said, Md. A. Md.

247

Session 9 - Earth Explorer Coal fire quantification and detection using the DLR experimental Bi-Spectral Infrared Detection (BIRD) small satellite Tetzlaff, Α., Zhukov, B„ Himer, Α., Kuenzer, C., Voigt, S.

255

The STEAM project van Scheele, F., Frisk, U., Veldman, S.

263

NASA small satellites for exploration of the Earth-Sun system Neeck, S. P., Gay C. J.

269

Session 10 - Orbit & Attitude Determination Orbit analysis for the Brazilian-German space project MAPSAR based on user requirements Jochim, F., Quintino da Silva, Μ. Μ., Schröder, R., Hajnsek, I„ Puls, J.

275

Analysis of orbit propagation and relative position accuracy of small satellites for SAR interferometry De Florio, S„ Neff, T„ Zehetbauer, T.

284

Orbit analysis of a remote sensing satellite for local observation of the earth surface Ebrahimi, Α., Mirshams, M., Moosavian, S. A. A.

291

The Micro Advanced Stellar Compass for ESA's PROBA 2 mission Jorgensen, P. S., Jergensen, J. L„ Denver, T., van den Braembuche, P.

299

Spacecraft formation control in vicinity of libration points using solar sails Novikov. D.. Nazirov. R.. Eismont, N.

304

Session 11 - Subsystems Dependable and flexible board computer software for pico satellites Montenegro, S., Brieß, K„ Kayal, H,

312

Smallsat Propulsion Smith, P., McLellan, R„ Edwards, S„ Gibbon, D.

319

Centralised computation service architecture for the X-SAT micro-satellite Ramesh, B., Mohan, D„ Bretschneider, T., McLoughlin, I.V.

329

Session 12 - Attitude Control Magnetic attitude control systems of the nanosatellite 7'W-Series Ovchinnikov, M. Yu., Penkov, V. 1., Ilyin, Α. Α., Selivanov, S. A.

337

Analysis of the HAUSAT-2 attitude control with a pitch bias momentum system Chang, Y.-K., Lee, B.-H., Kim, S.-J.

345

A micro sun sensor for earth-observation nanosatellites flying in formation Rufino, G., Grassi, M., Pulcino, V., Degtyarev, A.

353

Experiences with the BIRD magnetic coil system for wheel desaturation, satellite rate damping and attitude control - the concept of virtual wheel systems Terzibaschian. T. Spacecraft combined attitude and thermal control system Varatliarajoo, R., Fasoulas, S.

361

369

Session 13 - Data Transmission & Ground Stations Secured communication links for earth observation satellites Michalik, Η.. Hinsenkamp, L.. von der Wall. M., Burr, T., Schönenberg, A.

377

BAYERNSAT - a real-time communication architecture for interactive earth observation via geostationary inter satellite link for small satellite systems Letschnik, J., Raif, M., Pauly, K., Walter, U.

385

A web-based modular and flexible data acquisition and telemetry monitoring system for micro satellites Sarkarati, M„ Brieß, K„ Kayal, H.

392

Space-based AIS for global maritime traffic monitoring Heye. G. K., Eriksen. T„ Meland, B. J., Narheim. Β. T.

401

Session 14 - Technology Demonstration Pico satellite concept of TU Berlin Kayal, H„ Brieß, Κ.

410

Technology demonstration with the micro-satellite Flying Laptop Grillmayer, G., Falke, Α., Röser, H.-P.

419

ΝΕΑ detection, a possible use of the flying laptop micro-satellite reconfigurability Jorgensen, J. L., Jorgensen, P. S., Grillmayer, G.

428

Session 15 - Disaster Monitoring Evaluation of the utility of the disaster monitoring constellation in support of earth observation applications Underwood. C. I., Mackin, S., Stephens. P., Hodgson, D„ da Silva Curiel, Α., Sweeting, M. The micro-satellite "Chibis" - universal platform for development of methods of space monitoring of potentially dangerous and catastrophic phenomena Zelenyi, L. M., Rodin, V. G., Angarov, V. N., Breus, Τ. K., Dobriyan, Μ. B., Klimov, S. I., Korablev, Ο. I., Korepanov, V. E., Linkin, V. Μ., Loupian, Ε. Α., Ivanov, Ν. Ν., Lopatento, L. Ε., Sedykh, Ο. Yu. Small satellite utilization for disaster management information dystems Albayrak, O.

436

443

452

PREFACE The 5th IAA Symposium on Small Satellites for Earth Observation, initiated by the IAA Committee on Small Satellite Missions, was again hosted by DLR, the German Aerospace Center, with its Berlin-Adlershof site. The participation of scientists, engineers, and managers from 24 countries reflected the high interest in the use of small satellites for dedicated missions applied to Earth observation, from scientific Earth observation missions to technology demonstration missions. Out of more than 120 paper submissions in response to the Call for Papers, 62 candidates had the possibility to present their papers orally. Thirty-three participants made use of the opportunity to present their paper proposal as a poster. As in the preceding symposia, the contributions showed that dedicated Earth observation missions cover a wide range of very different tasks. These missions provide increased opportunities for access to space and can be conducted relatively quickly and inexpensively. The spacecraft bus, the instruments, and the ground systems can be based either on optimized off-the-shelf systems with little or no requirements for new technology, or on new high-technology designs. Thus a new class of advanced small satellite missions, including autonomously-operating "intelligent" satellites and satellite constellations can be created, opening new fields of application for science and the public. The symposium provided 15 sessions for oral presentations, one poster session, two panel discussions and several social events. The two panel discussions were dedicated to "Small satellites contributions to disaster warning and mitigation" and to "Costeffective Earth observation missions". Furthermore, in our 5th Symposium the Student Prize Paper Competition has been continued. The student papers have been evaluated by distinguished judges selected from academia, industry and government, coming from four continents. The finalists presented their papers in the Student Conference session. With all these events the symposium offered many opportunities for exchanging information, exploring new concepts, and developing new collaborative relationships among individuals and institutions, industry and academia. These outcomes of the symposium complement and reinforce the purpose of the Academy. All contributions (oral and poster contributions) of the 5th symposium were published in their short 4-page version in the symposium digest (Small Satellites for Earth Observation, Digest of the 5th International Symposium of the International Academy of Astronautics, Berlin, April 4-8, 2005, ISBN 3-89685-570-0, 566 pages). This book of selected proceedings contains only orally presented papers and a small number of poster contributions the long versions of which have been made available to the editors within a given time frame after the symposium. We would like to thank the members of the Scientific Program Committee and the Program Committee for their active support in organizing the Symposium and selecting the papers to be presented orally and as posters, as well as the panel chairs and panelists who all substantially contributed to the well-perceived program. We are also grateful to

the session's chairpersons and rapporteurs. Special thanks go to the Symposium and Program Coordinator, Stefan Röser assisted by Ute Dombrowski and Karl-Heinz Degen, without whose efforts the organization of the symposium and of the proceedings would not have been possible, and to the Symposium Chief Rapporteur, Larry Paxton, who assembled and condensed the summaries provided by each session's rapporteur into a coherent symposium synopsis.

The Symposium Chairmen

Arnoldo Valenzuela

Hans-Peter Röser

Rainer Sandau

SMOS: AN L-BAND INTERFEROMETRIC RADIOMETER MISSION F. Bermudo (1) , M. Venet (1) , M. Le Du(1), P. Landiech (1) , Y. Kerr (2) , H. Barre(3) (1)

CNES, Toulouse. France, [email protected]

[email protected], [email protected]. [email protected] (2) (3)

CESBIO. Toulouse, France, [email protected]

ESTEC, Noordwijk, The Netherlands, [email protected]

ABSTRACT The SMOS (Soil Moisture and Ocean Salinity) mission is a joint ESA/CNES/CDTI* Earth Observation program. The SMOS mission has been selected as the 2nd Earth Explorer Opportunity Mission with a tentative launch in September 2007. The objective of the SMOS mission is to provide Soil Moisture (SM) and Ocean Salinity (OS) maps. Both SM and OS are key variables in climate monitoring, surface/vegetation/atmosphere transfers, and ocean/atmosphere cycles. The SMOS spacecraft will consist in a platform, based on the CNES / ALCATEL PROTEUS generic platform adapted to the mission specificities and a payload, procured by ESA. The payload is an EADS/CASA L-Band (1.4 GHz) 2D passive interferometric radiometer with a Y-shaped 3 arms synthetic aperture antenna. The SMOS Satellite Operations Ground Segment (SOGS), in charge of operating and controlling the satellite, will be adapted by CNES from generic PROTEUS ground segment and its Command & Control Centre is installed in Toulouse within the CNES premises. A SMOS Data Processing Ground Segment (DPGS), devoted to acquire, process, and dispatch the SMOS Levels 1 and 2 products, is being developed and located in the ESA/Villafranca centre. A Centre Aval de Traitement des Donnees SMOS (CATDS) dedicated to process, archive and dispatch the SMOS scientific data at levels 3 and 4 is being developed by CNES. An overview of the SMOS mission will be addressed by this paper with attention put on the CNES specific contribution in the frame of this cooperation.

1

MISSION OBJECTIVES AND MEASUREMENT APPROACH

Soil Moisture and Ocean Salinity data are urgently required for hydrological studies and for improving our understanding of ocean circulation patterns. The importance of estimating soil moisture in the root zone is paramount for improving short- and medium-term meteorological modelling, hydrological modelling, the monitoring of plant growth, as well as contributing to the forecasting of hazardous events such as floods. The amount of water held in soil, is of course, crucial for primary

*

ESA - European Space Agency, CNES - Centre National d'Etudes Spatiales, CDTI - Centro para el Desarrollo Technologico Industrial

7

production but it is also intrinsically linked to our weather and climate. This is because soil moisture is a key variable controlling the exchange of water and heat energy between the land the atmosphere. Precipitation, soil moisture, percolation, run-off, evaporation from the soil, and plant transpiration are all components of the terrestrial part of the water cycle. There is, therefore, a direct link between soil moisture and atmospheric humidity because dry soil contributes little or no moisture to the atmosphere and saturated soil contributes a lot. Moreover, since soil moisture is linked to evaporation it is also important in governing the distribution of heat flux from the land to the atmosphere so that areas of high soil. In the surface waters of the oceans, temperature and salinity alone control the density of seawater - the colder and saltier the water, the denser it is. As water evaporates from the ocean, the salinity increases and the surface layer becomes denser. In contrast. precipitation results in reduced density, and stratification of the ocean. The processes of seawater freezing and melting are also responsible for increasing and decreasing the salinity of the polar oceans, respectively. As sea-ice forms during winter, the freezing process extracts fresh water in the form of ice, leaving behind dense, cold, salty surface water. If the density of the surface layer of seawater is increased sufficiently, the water column becomes gravitationally unstable and the denser water sinks. This process is a key to the temperatureand salinity-driven global ocean circulation. This conveyor-belt-like circulation is an important component of the Earth's heat engine, and crucial in regulating the weather and climate. Another important aspect of this mission is to demonstrate a new measuring technique by adopting a completely different approach in the field of observing the Earth from space. A novel instrument has been developed that is capable of observing both soil moisture and ocean salinity by capturing images of emitted microwave radiation between 1400-1427 MHz (L-band) : SMOS, with launch in September 2007. will carry the first-ever, polar-orbiting, space-borne, 2-D interferometric radiometer. 1.1 Retrieval of Sea Surface Salinity from L-band microwave radiometry Salinity describes the concentration of dissolved salts in water. It measures in practical salinity units (psu), which expresses a conductivity ratio. The average salinity of the oceans is 35 psu, which is equivalent to 35 grams of salt in 1 litre of water. SMOS will

8

aim at observing salinity down to 0.1 psu (averaged over 10-30 days and an area of 200 km χ 200 km) - which is about the same as detecting 0.1 gram of salt in a litre of water. The sensitivity of L-band (1.4 GHz) passive measurements of oceanic brightness temperature TB to SSS is well established. The dielectric constant for seawater depends on both SSS and SST. So, it is possible to obtain SSS information from L-band passive microwave measurements if the other factors influencing TB can be accounted for. The GODAE (Global Ocean Data Assimilation Experiment) optimised requirement for Open Ocean SSS is 0.1 psu over 200 km boxes over 10 days. This GODAE requirement can be obtained by averaging the SMOS individual measurements in both space and time. The averaging procedure requests excellent stability and calibration of the radiometer receiver. Factors that influence TB in addition to SSS, and are to be corrected for include Sea Surface Temperature (SST), surface roughness, foam, sun glint, rain, ionospheric effects and galactic/cosmic background radiation. Estimates for the uncertainties associated with some of these have been made: use of L-band radiometiy for the measurement of SSS from aircraft has been demonstrated using a combination of modelling and ancillary data on SST and wind speed. 1.2 Retrieval of SM from L-band microwave radiometry Moisture is a measure of the amount of water within a given volume of material and can be expressed as a percentage. From space, the SMOS instrument will measure with a sensitivity of 4% the volumetric moisture over the surface of the Earth - this amount is about the same as being able to detect less than one teaspoonful of water mixed into a handful of dry soil - with a spatial resolution of 35-50 km. Due to the large dielectric contrast between dry soil and water, the soil emissivity at a microwave frequency F depends upon moisture content. The vegetation cover attenuates soil emission and adds a contribution to the radiation temperature TB. However, at L-band, this attenuation is moderate; TB is sensitive to soil moisture for vegetated areas with biomass < 5 kg m-2. Previous research has shown the strong advantages of L-band microwave radiometry for measuring surface soil moisture. At higher frequencies, vegetation attenuation increases, and a much smaller portion of the earth would be accessible. Passive microwave techniques have all-weather capabilities, the signal-tonoise ratio from dry to wet soils is significantly higher for a radiometer than for radar. Furthermore, the signal is less sensitive to structural features of the surface such as soil roughness or canopy geometry. The brightness temperature TB depends on three surface variables of interest: soil moisture, vegetation layer optical depth, effective surface temperature Ts (Κ). To discriminate between the effects of these factors, the radiometer offers the possibility of acquiring data for several polarisations Ρ (Η & V), and several incidence angles i (0° to about 55°).

2

PAYLOAD CONCEPT

In order to achieve the required spatial resolution for observing Soil Moisture and Ocean Salinity a huge antenna (about 8 m) would normally be necessary. For the SMOS

9

mission, however, the antenna aperture has been cleverly synthesized through a multitude of small antennae. After more than 10 years of research and development, with the aim of demonstrating key instrument performance such as antenna deployment and image validation, the innovative SMOS instrument, called MIRAS (Microwave Imaging Radiometer using Aperture Synthesis) has been realised. MIRAS consists of a central structure and three deployable arms (4 m deployed length), each of which has three segments. During launch, these arms are folded-up, and after separation from the launch vehicle and satellite attitude acquisition, the antenna arms are deployed via a system of spring-operated motors and speed regulators. The instrument enter then the operational phase with 3 different operating modes: calibration mode, dual polarisation mode (in this mode horizontal and vertical polarisation acquisitions correlations are interlaced in transmitted data), and full polarisation mode (in this mode, by alternating horizontal and vertical polarisation acquisitions of arms between them). There arc 69 antenna elements, called LICEF receivers, which are equally distributed over the three arms and the central structure called the HUB. Each LICEF is an antennareceived integrated unit that measures the radiation emitted from the Earth at L-band. The acquired signal is then transmitted to a central correlator unit, which performs interferometry cross-correlations of the signals between all possible combinations of receiver pairs. 3 calibration unit (Noise Injection Radiometers) based on a noise diode associated to a splitter and controlled by the CCU allow to calibrate each visibility function. The correlator data output are formatted and stored in a Solid State Memoiy housed in the CCU. The instrument acquisition data are finally transmitted to ground via an X-band channel. All units are accurately thermal controlled separately in each arm and in the HUB. to insure stable thermal conditions.

3

PROTEUS SYSTEM MAIN PERFORMANCES

The SMOS payload will be carried on a standard spacecraft bus called PROTEUS, which was developed by the French space agency (CNES) and Alcatel Space Industries. It is a generic multi-mission platform (utilized by JASON1-2001, CALIPSO-2005, COROT-2006...) with well-defined interfaces so that with limited adaptations, the

10

SMOS scientific instrument can be mounted on the top of the spacecraft. The PROTEUS system architecture consists not only in a generic platform, but also in the different required means to validate and integrate it (SW and Systems validation benches, AIT GSE), as well as in a generic ground segment. Interfaces for a dedicated mission are at satellite level between the Platform and the Payload, and at ground level between Satellite control centre and Mission control centre. 3.1 PROTEUS Bus The Bus design has been driven by such concerns as simplification of the system architecture - by suppressing the subsystem level and replacing it by functional chains directly merged at system level -, allowance to introduce commercial components with adapted EEE procurement and selection as far as possible of off the shelf equipment units. Although the reduced spacecraft bus size, measuring just one cubic metre, it acts as a service module accommodating all the subsystems that are required for the satellite functions. Main platform specifications are gathered in fig 3. Satellites based on PROTEUS are in the 500 kg class, with originally 200 kg and 150 W devoted to Payload. Post JASON1 modifications implementation allow now to face increased demands in the range of 370kg/350W for Payload. The electrical on-board command and data handling architecture is centralised on a Data Handling Unit, including a 3 Gbits mass memory storage capacity, plus a MIL 1553 bus dedicated to Payload Remote Terminals (RT). It manages the satellite operational modes and interfaces with the payload central processor unit, forwarding payload commands received from the ground and supplying all auxiliary satellite data that are needed by the payload to fulfil its scientific measurements. After launch, when the spacecraft has separated from the launcher, there will be an automatic start-up sequence, which will result in the deployment of 2 symmetrical solar arrays wing with classical silicon cells, generating the electrical power. The energy is stored in a lithium-ion battery and distributed through a non-regulated electrical bus. The platform thermalcontrol subsystem relies on passive radiators and active regulation by heaters. For SMOS, payload provides its own thermal regulation. PROTEUS uses a GPS receiver for orbit determination and control, which provides satellite position information, and a hydrazine monopropellant system for four 1-Newton thrusters that are mounted on the base of the spacecraft Nominal attitude control is based on a gyro-stellar concept. The autonomous Star Tracker type (lost in space acquisition capacity) which provides accurate attitude information for both the instrument measurements and the satellite attitude control, is accommodated on the payload to minimize thermo-elastics concerns. Three 2-axis gyroscopes are used to measure the change in the spacecraft orientation, and thus provide the accurate attitude knowledge needed to fulfil stability and pointing requirements. Four small reaction wheels generate torque for attitude adjustment. A quasi autonomous nominal mode (only guidance TCs, sent in advance for several days) with satellite modes number minimized, FDIR (Failure Detection and Recovery) principles completely autonomous and robust Safe Hold Mode- not limited in time- , allow to reduce drastically operations concept to working hours-working days only.

11

Mass

Platform: up to 300 kg, Payloads up to 350 kg

Orbits

any LEO from 20° to SSO, 700 km to 1500 km

Launchers

Any small SL launcher - associated fairing volume

Power

Up to 300 W Platform + 350 W Payload

Orbital capacity

Up to 120 m/s for a 500 kg spacecraft

Autonomy

Ability wrt 1 ground station, operated during working days only

Pointing

Any pointing

Pointing accuracy

0.05° bias + 0.05° 3σ per axis + 1 .E-3°/s low frequency stability

Payload IF

Dedicated MIL STD 1553 bus 160 kbps + dedicated TM/TC

Data storage

500 Mbits Bus + 2 Gbits Payload, End of life

TMTC

S band, 800 kbps TM, 4 kbps TC, CCSDS

Orbitography

Use of an on board GPS receiver: no ground station angular measurements, nor ranging,

Lifetime

3 years. All elements subject to aging + radiations sized to 5 years Figure 3: PROTEUS PF main performances

3.2

Satellite Operations Ground Segment (SOGS)

SOGS will be derived from the Generic PROTEUS Control Centre architecture (cf.fig.4) which was developed in parallel with the platform concept. It consists in several TMTC earth terminals (TTC-ET), a control centre (CCC) and a data transmission network (DCN). Mission Centre (MC) interfaces with CCC and if needed also directly with TTC-ETs via the DCN, for fast Payload data retrieval. To minimize costs, off the shelf products and standards were systematically considered. As well, TTC-ETs were specified not to require any operator, remotely commanded from CCC and monitored through TM Remote Monitoring (RM). CNES network of several TTCETs called ICONES has been built and will be shared by both PROTEUS based missions and MYRIADES micro-satellites ones. CCC architecture is organized around 4 main functions: Gl-TC generation and TM retrieval; G2-Orbit restitution, TTCETs ephemeris. guidance TCs; G3-TM storage, linked to an intranet server; G4-Mission centre interface. All routine operations have been defined so as to be autonomous and managed by an Agenda procedure, except TCs sending which requires operator presence. After each path, housekeeping TM is filtered through ground surveillances. As well, each agenda task gives a status of execution. A warning surveillance subsystem monitors in parallel logbooks and is able to warn operators. Based on commercial PCs, this CCC may be easily duplicated at low costs.

12

CCC, located in Toulouse CNES centre, will operate the spacecraft via an S-band station in Kiruna. Sweden and interface with the Payload operations Programming Centre (PLPC), located in ESA-VILSPA facility in Villafranca, in charge of monitoring, controlling and programming the operations of the SMOS Payload.

4

MISSION PROCESSING CENTERS

The SMOS instrument data products after calibration and corrections performed on ground shall be maps of brightness temperature at different polarizations for all specified land areas and for the oceans. Subsequently from these data, geophysical information of concern to the mission, namely fields of Soil Moisture and Ocean Salinity, will be extracted by mission processing centres: an ESA/CDTI Level 1 and 2 Data Processing Ground Segment (DPGS) bi-located at the ESA-VILSPA facility in Villafranca and the ESRIN facility in Frascati. The DPGS, in charge of acquiring, processing, archiving and dispatching the SMOS scientific data is mainly composed of: an X-Band Acquisition Station, a Payload Data Processing Centre which main function is to process, calibrate and archive the SMOS scientific data up to level 2 and a SMOS User Service centre insuring interfaces and services between the SMOS System and the external users. - a CNES "Centre Aval de Traitement des Donnees SMOS" (CATDS) in charge of SMOS scientific data at levels 3 and 4 included. The CATDS will process, calibrate, archive and dispatch the SMOS scientific data at levels 3 and 4 derived from the Levels 1 and 2 products. The CATDS architecture is based on two main types of components (cf. fig. 5): two Expertise Centre (OS and SM) which will host the development, validation and improvement of the L3 and L4 algorithms in close co-operation with the scientific

13

community, which will assess the quality of the products and will provide specific information to users. one single Processing Data Centre (PDC), which will routinely produce and disseminate L3 and L4 data from auxiliary data and from LIB, which is the last product level not separated on the basis of land or sea, provided by DPGS. The option considered is to not decorrelate the processing of the two kinds of products: CATDS will produce simultaneously OS and SM products. L3 and L4 products will be produced and disseminated once per day. They are based on the last 10 days of LIB. PDC will also manage products catalogue and archive facility. The CATDS will also probably help to fine tune the calibration for SMOS algorithms, improve and validate algorithms from LO to L2 processing. Once validated, these algorithms will be proposed for a transfer to DPGS.

Figure 5: CATDS architecture

14

EARTH OBSERVATION - A NEW PARADIGM Prof Sias Mostert+*, Prof Jan du Plessis *, Thys Cronje * +University of Stellenbosch and *SunSpace, Stellenbosch, South Africa [email protected] ABSTRACT The impact of disruptive technologies, allows for a completely new paradigm for the use of earth observation satellites. Improvements in sensor technologies and the ready availability of LEO launch access, creates the platform for a new paradigm for remote sensing. Instead of expensive single platform satellites of which all the image data is archived, a new generation of remote sensing satellite provides such an effective cost/performance ratio, that only the images that is application specific need to be tasked. In this paper, the impacts of the disruptive technologies, will be explored in the context of 1) The Multi-Sensor Micro-satellite Imager (MSMI) next generation imaging payload for earth observation currently under development and 2) Far smaller satellites that have similar spectral and ground sampling distance as larger imager payloads on larger satellites 1. Introduction Recent advances in Commercial-Off-the-Shelf (COTS) sensor- and storage technology is enabling a completely new class of micro-satellites. Ground Sampling Distances (GSD) smaller than 5m. that was only possible on larger satellites, are now possible on satellites with a mass of less than 60kg due to smaller pixel sizes, refractive optics and accurate ADCS pointing and viewing control. The significantly smaller satellites reduces the total cost of launch deployment for a number of satellites, making it economical to deploy a constellation of interactively controlled micro-satellites for the same cost as a single micro-satellite in the $15M class. Two of the implications are 1) a redundancy of available space resources that drives a new operational paradigm and 2) high resolution imagery with a temporal resolution that allows time-varying processes to be monitored. The MMSat satellite, that forms the baseline of the ZASat 1 Pathfinder mission combines the current advances in sensor and digital storage technology to create a 6.5m GSD satellite, scheduled control or interactively controlled, with 6 spectral bands and up to hundreds of Gigabytes storage capacity in a satellite of 60kg. Utilising this next generation of technology, leads to a matrix camera satellite of less than 40 kg with a snapshot imaging capability smaller than 5m GSD. Although 5m to 30m GSD multi-spectral remote sensing from micro-satellites has been demonstrated a number of times the cost basis of around $10M to $15M per mission has

15

precluded wide scale adoption of such micro-satellites for personal observation functionality as well as high temporal resolution systems, where four or more satellites are required in constellation. Furthermore, a ground sampling distance of smaller than 10m is often required for resource management, while for infrastructure assessment a resolution of smaller than 5m is required. The need is to lower the cost of satellites by a factor of 5 to 10 times before a constellation of satellites with a combined swath of 600km can be achieved at a ground sampling distance of 5m GSD. Existing EO micro-satellites are mostly characterized by single satellites demonstrating a specific capability. The exceptions are the DMC [1] and RapidEye [2] based in Germany, as well as the planned ARM (African Resource Management) [3] constellation. This paper reviews the expected impact of two technologies that have a disruptive impact on remote sensing satellites. The next section reviews some of the paradigm shifts that result from the technology advances. The impacts of the technology advances on micro-satellite solutions currently under development are explored, before closing the paper with an assessment of the impact of the technology innovation on high temporal resolution systems. 2. Disruptive Technology The miniaturisation in sensor and storage technology enables two innovations in microsatellite technology: a) Smaller optical systems are required due to small CCD sensor pixel sizes that allow for a resulting smaller optical system for a specific Ground Sampling Distance. b) Storage technology of hundreds of Gigabytes opens completely new avenues for onboard image buffering and processing. The increase in storage capacity per unit volume, enables new operational paradigms for data access, ie. storage of images on-board for far longer due to increased capacity. Further more, since the downlink is bandwidth limited, this provides the possibility to increase the on-board processing capability to improve the information extraction and hence lower the required data rate to the ground for quick reaction data, while still buffering all the raw data on-board, for in case the raw data has to be downloaded later. Initial investigations in on-board processing for example to extract Ground Control Points, is ongoing. Other expected impacts of the new generation of technology are: 1. Flying different single sensors on different platforms to create a non-heterogeneous constellation 2. Combining large platforms with particular sensor suites with a constellation of smaller satellites with a smaller, but broader overview sensors

16

3. On-board processing for information extraction and transmitting only the relevant information, while increased on-board storage allows for buffering the original unprocessed data for later downloads if needed. 4. Distributed ground stations, where multiple access to the satellite constellation with viewfinding bore point selection as pioneered by Prof Udo Renner of TUB with DLRTubsat The promise of smaller optical systems would lead to smaller satellites that would increase the cost/performance ratio of satellites to the point where the mission specific aspects can be fine tuned to a local geographic area. The next section examines some of the implications of the smaller sensor size technology on imager performance by using two examples of projects currently under development. 3. Smaller Optical Systems The impacts of smaller sensors are examined on the design constraints of imaging systems, as it relates to the performance in three areas of optical instruments viz. spatial resolution, radiometric resolution and swath width. 3.1 Spatial Resolution A decrease in pixel pitch of newly available CCDs, will result in a smaller F# optical system to keep the performance detector limited and thus increasing the constraints on the optical system. The Aiiy disk diameter D caused by diffraction is one of the important parameters which can be related to the detector pixel pitch, but does not describe the image quality from a spatial resolution point of view: D = 2.44 λ F# For a pixel size of 5μηι an optical system with F# of less than 3.7 is needed according to the above criteria. It will be shown that optical systems with a larger F# could produce sufficient modulation for good image quality at high spatial frequencies. The increased performance requirements will however not only increase the constraints on the optical design but also on the Focal Plane Assembly (FPA). The relative alignment of the detectors will also become crucial, due to the fact that the focal depth of the optical system will decrease with the reduction in pixel pitch and F#.

17

3.2 Radiometric Resolution (SNR) The decrease in GSD will also result in a decrease in integration time on a CCD pixel, to a point where it is too short for sufficient good signal and signal to noise ration (SNR). To overcome this obstacle Forward Motion Compensation (FMC) could be used to decrease the ground track velocity and thus increasing the integration time to produce target Noise Equivalent Ratio (NER) values of less than 0.5%. A shorter integration time will result in the faster clocking of the detector and thus an increase in current consumption (I = C dV/dt and for CCDs the capacitive load is relatively large). Close attention must be given to the thermal design of the FPA. 3.3 Swath Another parameter to be optimized is the swath width of the sensor and will determine, with the GSD, the number of pixels needed in the across track. The swath could be increase by either placing more than one imaging system, aligned to an angle with each other, on the same platform or by increasing the field size of the optical system through the use of a refractive optical system.. 3.4 Impact of Disruptive Technology on Imaging Systems in Development The impact of the availability of CCDs with a pixel pitch of smaller than 10 microns can best be seen by comparing the design objectives of the MSMI imager with the next generation MMSat imager. 3.4.1 The MSMI Imaging Payload The MSMI payload combines a unique set of sensors consisting of a panchromatic sensor aimed at infrastructure assessment, a multi-spectral payload sensor for natural resource management and a hyper-spectral sensor for vegetation process modeling research. The MSMI payload design drivers are reviewed, followed by Table 2 comparing the specifications with the MMSat imager. One of the important aspects to note is the significantly small optical system on MMSat that provides a comparable multi-spectral performance to MSMI, but with a lower total data volume. The challenges for the Multi Sensor Micro-satellite Image are to meet the desired performance requirements over the very wide spectral range and to integrate all the sensors (Panchromatic, Multi-Spectral and Hyper-Spectral) on a single Focal Plane Assembly (FPA). A reflective configuration was chosen because of symmetry, inherent insensitivity to air pressure and thermal gradients, spectral width, the physical size of the objective and the ease to assemble and test. Due to the large number of sensors in the focal plane a relatively large field size was needed. Therefore the major trade-offs were obscuration ratio and back focal length vs.

18

MTF performance and transmission. To be able to accommodate all the sensors fold mirrors were used to fold out the field to the respective detectors. The diffraction limited MTF is the ultimate performance measurement to determine the quality of the imager. By analysing the through focus MTF curves the amount of astigmatism could also be determined. Table 1 summarises the expected MTF values.

0.0 30.9%

Expected Optical MTF values after Tolerances MTF @ 72 cy/mm MTF (a), 42 cy/mm 0.7 1.0 0.0 0.7 1.0 49.4% 22.7% 17.3% 43.7% 34.9% Table 1.

The design optical MTF of MSMI

3.4.2 Comparing two Imagers The impact of sensors with a finer pixel pitch, enables the possibility in principle of utilizing far smaller optics. The following table compares the specifications of the MSMI imager with the MMSat imager, in order to show the kind of performance that can be expected from a far smaller imager that would enable smaller satellites and hence a less expensive deployment cost.

Aperture Dia. Focal Length F# Optical Design Instruments Pixel Pitch # Spectral Bands Spectral Range MTF (a), Nyq. Intended Orbit GSD Swath

MMSat 80mm 400mm F5 Refractive MS 5μηι 6 440 - 900nm 40% 500km 6.25m 46km

MSMI 280mm 1.72m F6.14 Reflective Pan 7μηι 1 440 - 685nm 23% 2.7m 23km

MS 12μιη 6 440 - 900nm 35% 660km 4.6m 27km

HS 38μηι 200 0.4 - 2.35μιη 30% 15m 15km

Pan - Panchromatic, MS - Multi-Spectral, HS - Hyper-Spectral Table 2. Two imaging systems with impacts of new technology. Given the technology of smaller imagers can be utilized successfully, one can expect an abundance of remote-sensing satellites in space due to a lower cost to access space. If remote sensing satellites are in abundance, new paradigms are possible for the application of remote sensing. The next section explores some of the expected impacts. 4. New Paradigm for Remote Sensing The improvements in technology leading to far smaller imaging satellite and larger onboard storage, , opens new ways of creating and applying remote sensing satellite

19

systems. Some of the potential implications of the technology advances are listed as it relates to remote sensing satellites. 1. The lower cost of remote sensing should lead to new remote sensing applications with a limited geographic area with an imaging area that is local to a specific geographic area. 2. A 3 to 5 times less expensive space resource allows the tasking of data sets only when they are required, without having to store all possible usable data in the future by archiving all collected data sets. 3. The construction of heterogeneous high temporal resolution constellations is possible that includes satellites with a high spatial resolution as well as satellites with high spectral resolution. 4. The improvements in storage and processing capacity that is suitable for flying on micro-satellites allows far more data to be collected, buffered and processed before providing the output as processed information to bandwidth limited ground-stations. 5. Dedicated imaging processor with interactive viewfinder capability enables cloud avoidance and hence resource saving by minimising the non-usable data collected. 6. Steering the satellite by means of an interactive viewfinder system to advantageous areas for multispectral or hyperspectral line-scanning imaging opens a mission-operation mode that complements the standard scheduled lists. The viewfinder can be controlled from multiple transportable TT&C containerbased ground stations. Improvements in Commercial Of The Shelf (COTS) technology for ground-stations, are also impacting significantly on the access to remote sensing data and information. Some of the impacts on the utilization of a space based observation system are: The satellite collected information can be directly transferred to low cost end user terminals, eliminating the need for high bandwidth telecommunication infrastructure. Portable ground stations that can be moved to where the information is required. Distributed satellite management and information distribution. The outstanding issue of the impact of the disruptive technology, relates to the cost/performance improvements. The next section explores some of the assumptions and predicts cost targets that would unlock the potential of utilizing micro-satellites on a grand scale. 5. High Temporal Resolution Cost/Performance Implications The inclusion of the technology advances in a system of remote sensing satellites, yields a completely new solution to high spatial- and high temporal resolution remote sensing.

20

It is however important to evaluate whether enough progress has been made in the cost/performance ratio to enable the cost effective deployment of a satellite constellation. Increasing the temporal resolution to a daily revisit time at a ground resolution of 5m, requires a constellation of 4 to 5 satellites. At a typical system cost of $ 15M, this limits the deployed cost per spacecraft to $3M. This cost level is approximated with the improved sensor, processing and storage capability of the MMSat class of satellites. A system that provides a three times daily revisit time requires the use of 12 satellites in three constellation planes. With a system cost constraint of $ 18M the deployed cost per spacecraft is required to be only $1M5. The objective would be to observe any location on the globe three times a day, as is approached with the cost of the MxSat class of satellites. 6. Expected Results Some of the results expected from the first MMSat mission, is an operational remote sensing satellite that will produce data sets at a ground sampling of 6.5m GSD from a 60kg satellite. Compared to Sunsat 1, the MMSat mission GSD would be more than two times smaller the Sunsat 1 ground sampling distance, while maintaining the swath width and twice the number of spectral bands and with a 24 Gigabyte storage capacity. Current technology programs indicate improvements of storage 10 times more that is currently available. The results expected from the first MSMI mission, is the first hyper-spectral information at a Ground Spacing Distance of 15m from a satellite in the mass range of 200kg. This data is over and above the multi-spectral and panchromatic data that is to be applied in the African Resource Management (ARM) constellation. 7. Conclusion An abundance of space based remote sensing resources is one of the expected outcomes of the technology advances in COTS sensor and digital storage technology. New paradigms for the application of remote sensing are expected to develop as a result of the possibility of having a personal remote sensing resource to complement more conventional Earth resource large satellites; much as the personal computer opened up the access to computing for a larger user community without replacing mainframes, but complementing them. Note: The MSMI payload is a project that is jointly funded by the Innovation Fund in South Africa and the Flemish Government and executed by SunSpace, CSIR and Stellenbosch University in South Africa and Catholic University of Leuven and OIP in Belgium. 8. References

21

[1] Underwood, C.I. et al, Evaluation of the Utility of the Disaster Monitoring Constellation in Support of Earth Observation Applications, IAA Small Satellites for Earth Observation, IAA-B5-1501, Berlin, April 2005. [2] Neeck, S.P. et al, NASA Small Satellites for Exploration of the Earth-Sun System, IAA Small Satellites for Earth Observation, IAA-B5-904, Berlin, April 2005. [3] Mostert, S. et al. African Resource and Environmental Management Constellation, IAC 2004,, IAA-11-4, Bremen, October 2003.

22

BIRD Fire Recognition and Comparison with Terra/MODIS D. Oertel1, E. Lorenz 1 , Β. Zhukov 1,2 ,1. Csiszar3 'DLR, Optical Information Systems, D-12489 Berlin, Rutherfordstr. 2, Phone: +49 30 67055523, [email protected], 2 on leave from Space Research Institute of the Russian Academy of Sciences 3 University of Maryland, Department of Geography, College Park, MD, USA

ABSTRACT Data of the MODerate resolution Imaging Spectro-radiometers (MODIS) on the Terra and Aqua satellites are widely used for global and regional monitoring of active fires. MODIS, with their infrared channel resolution of 1 km at nadir do omit small fires, which leads to an underestimation of the cumulative fire radiative power (FRP) on a regional or global scale. In order to estimate the significance of this effect, we compared the number and characteristics of fires detected by Terra/MODIS with those detected by the DLR small satellite BIRD. BIRD is dedicated for fire recognition, providing a resolution of 370 m at nadir in its infrared channels and allows the detection of smaller fires than MODIS by a factor of 7. For the comparison we used data acquired nearsimultaneously by Terra/MODIS and BIRD over fires in Siberia, Australia and Portugal. The results show that (1) the FRP of more than a half of the hot clusters, which were detected by BIRD, is below the detection limit of MODIS, (2) MODIS does only slightly underestimate the cumulative FRP in ecosystems where large fires take place and therefore (3) MODIS is hardly suitable for an early fire detection, but it is an adequate instrument for cumulative FRP estimation. 1. INTRODUCTION The MODerate resolution Imaging Spectro-radiometers (MODIS) on NASA's Earth Observing System (EOS) Terra and Aqua satellites have been used semi-operationally for global fire detection and monitoring since 2000 and 2002. respectively. The MODIS instruments, which are providing global data for land, ocean, and atmosphere products, may be also considered as the "workhorses" of global space-borne fire recognition. The MODIS Level 2 products MOD14 (Terra) and MYD14 (Aqua) are fundamental for fire detection. They assign to each 1-km pixel of the MODIS swath one out of nine classes, including three fire classes of various levels of confidence (from low to high). The current MODIS fire algorithm is described in detail in [1]. The experimental Bi-spectral IR Detection (BIRD) mission of DLR successfully demonstrated the capability of compact IR push broom sensors for high resolution fire detection and quantitative analysis of high temperature events [2], as it was reported at the 4th IAA International Symposium on Small Satellites for Earth Observation [3]. The objective of this paper is to reveal the complementary potential of two types of Low Earth Orbiting (LEO) IR sensors for innovative and quantitative active fire recognition:

23

• Whisk broom type spectro-radiometers, such as MODIS, with short pixel dwell time, and • Push-broom type imagers, such as the main sensors of BIRD, offering longer pixel dwell time, which allows the implementation of dynamic range extension by double exposition of the IR channels with different integration time. The BIRD main sensors consist of the Hot Spot Recognition System (HSRS) and the Wide-Angle Optoelectronic Stereo Scanner (WAOSS-B) [2], BIRD/HSRS is described in detail in [4], [5], Table 1: Main characteristics of the MODIS and BIRD sensors used for fire recognition HSRS / WAOSS-B on BIRD MIR: 3.4- 4.2 μιη TIR: 8 . 5 - 9 . 3 μιη NIR: 0.84- 0.90 μιη

MIR channel saturation Spatial resolution at nadir

MODIS on EOS -Terra / Aqua MIR: 3 . 9 - 4.0 μηι TIR: 10.8-11.3 μηι RED: 0.62 - 0 . 6 7 μηι NIR: 0.84 - 0.88 μιη 500 Κ 1 km

Swath width

2330 Ian

190 Ion

Revisit time

4 times a day

Experimental imaging of selected areas

Spectral channels for fire detection

600 Κ 3 7 0 m / 185m

2. ENHANCED FIRE DETECTION AND PARAMETER ANALYSIS RESULTS MODIS and BIRD are the only sensors world wide with Middle IR (MIR) and Thermal IR (TIR) channels that do not saturate from major fire signals. This allows: • an efficient rejection of false alarms, • an innovative approach for the satellite-based assessment of the amount of fuel burned and the estimation of direct pyrogenic carbon release, based on the use of radiative energy flux over active fires, the Fire Radiative Power (FRP) or Fire Radiative Energy' release (FRE), as a measure of the intensity of burning, [8], [9], and • the retrieval of further quantitative fire characteristics such as the effective fire temperature TF and effective fire area A/·. The BIRD fire detection algorithm uses the three nadir channels in a sequential thresholding procedure - as outlined in [5] and described in detail in [6] - including the use of the bi-spectral Dozier technique [7]. The principal channels for fire detection both in the MODIS and BIRD fire detection algorithms are the MIR channels at 3.9μηι and 3.8μιη respectively. The RED and/or Near IR (NIR) channel data of MODIS and BIRD, respectively, are used for the rejection of false alarm from sun glint. The Thermal IR channel data are combined with MIR channel data for: • rejection of false alarm from wann surfaces , and • retrieval of TF and AF in the sub-pixel domain, using the Dozier technique.

24

The FRP. a useful parameter for the characterization of the amount of burnt vegetation and of gas and aerosol emissions by a fire, can be assessed [9]: • using the effective fire temperature TF and effective fire area /I/.-, or • the MIR channel signal only (with sufficient accuracy for flaming fires with temperatures > 700 K). In this study the quantitative fire parameters were retrieved from the BIRD and MODIS data using the same methodology in order to provide directly comparable results. The methodology included the following steps: • consolidation of contiguous hot pixels in hot clusters, • retrieval of the effective fire temperature, effective fire area and FRP for the hot clusters, • estimation of the TIR background temperature variability (background clutter) and of the confidence intervals for the effective fire temperature, effective fire area and FRP, • and for pronounced fire fronts: estimation of the front length and radiative intensity (ratio of FRP to the front length, characterising the fire front strength). Fragments of the MODIS and BIRD fire scenes in Siberia, Portugal and Australia are shown in Figs. 1, 3 and 5. Histograms of the effective fire area and of FRP as estimated from the entire MODIS and BIRD image swaths are shown in Fig. 2, 4 and 6. The number of detected hot clusters and the total FRP in the fire scenes are compared in Table 2. Table 2. Number of detected hot clusters and their cumulative fire radiative power from the BIRD and MODIS data (the error intervals for FRP are indicated in brackets) Scene

Number of hot clusters BIRD MODIS

Cumulative FRP, Gigawatt BIRD MODIS

Angara, Russia, 12 June 2003

15

j

0.57 (0.47-0.57)

0.59 (0.37-0.59)

Angara, Russia, 10 July 2003

148

59

16.4(15.5-16.4)

14.3 (12.3-14.3)

Baikal, Russia, 16 July 2003

162

59

11.5 (10.7-11.5)

11.9 (10.7-12.0)

Portugal, 4 August 2003

99

35

15.5 (15.0-15.8)

12.0 (10.8-12.1)

Australia, 5 January 2002

227

34

5.2 (5.1-5.3)

2.9 (2.6-3.0)

BIRD allows much more detailed fire mapping, which manifests itself in a much larger number of detected hot clusters than the number of hot clusters in the MODIS data. BIRD could detect fires with FRP starting from ~1 MW, while the detection limit for MODIS was - 1 0 MW - see Figs. 2, 4 and 6 (here we characterise the fire detection potential of a sensor by the minimal detectable FRP that is much less sensitive to the fire temperature than the minimal detectable fire area). In this study only MODIS fire pixels with "nominal" and "high" confidence were considered. The contribution of the weak "low confidence" MODIS fire pixels to the cumulative regional FRP was negligible. In the MODIS data, some of small fires detected by BIRD are missed while others can not be separated from each other and are merged in larger clusters. Only the first effect (omission of small fires) leads to total

25

scene FRP underestimation from the M O D I S data. The magnitude of FRP underestimation by M O D I S in comparison to BIRD depends on the proportion of small fires in the scene, which in turn depends both on the ecosystem and on fire intensity. The strongest FRP underestimation of a factor of 1.8 was observed for the Australian bush fires, which are characterised by a relatively small fire front depth and as a consequence by a relatively small fire proportion in the pixel signal. On the contrary, forest fires in Siberia, which are expected to have wider fronts and therefore a larger fire contribution to the pixel signal, show a relatively small FRP underestimation by M O D I S of only 4 % for the three scenes. The 30 % underestimation of FRP by M O D I S for the forest fires in Portugal is between the corresponding values for the Siberian and Australian fires. An essential requirement for a sensor used for operational fire detection and monitoring is the capability to recognize fire fronts and estimate their characteristics, in particular the front radiative intensity. The comparison of M O D I S and BIRD fire scenes at Baikal (Fig. 1) and in Australia (Fig. 5) shows that M O D I S essentially allows the detection of only separate hot pixels and does not actually reveal clear fire fronts. In contrast, the higher spatial resolution BIRD data make it possible to recognise fire fronts and to estimate their characteristics (Table 3 ) : T F , Λ, . FRP, the front length, radiative front intensity and the effective depth (ratio of the effective fire area to the front length). The effective fire temperature was also used to determine the flaming ratio in the fire front (the ratio of the area of the flaming component to the area of the flaming and smouldering components), assuming that the flaming and smouldering components have a temperature of 1000 and 600 Κ respectively. Table 3. Characteristics of selected forest fire fronts in the BIRD image of the Lake Baikal area in Siberia, Russia obtained on 16 July 2003 (the error intervals for the fire parameters are indicated in brackets) Fire front in Fig.l 1 2 3 4 5 6 7

26

Efffire temperature, Κ

Flaming ratio

Effi fire area, Ha

FRP, MW

Front length, km 8.2

Front radiative intensity, kW/m 223

Front effective depth, m 7.7

851 (800-920) 711 (668-771) 775 (716-868) 783 (740-839) 850 (771-988) 860 (819-913) 763 (694-882)

0.30 (0.20-0.51) 0.08 (0.05-0.15) 0.16 (0.09-0.34) 0.17 (0.12-0.27) 0.30 (0.15-0.89) 0.32 (0.23-0.48) 0.14 (0.07-0.38)

6.3 (4.4-8.4) 1.1 (0.7-1.5) 2.1 (1.2-3.1) 0.53 (0.38-0.71) 0.43 (0.23-0.70) 1.9 (1.4-2.3) 0.73 (1.21-0.36)

1829 (1771-1829) 150 (136-150) 409 (377-409) 1 11 (105-111) 126 (121-126) 568 (554-568) 136 (123-136)

5.8

26

1.9

6.5

63

3.2

4.8

23

1.1

3.4

37

1.3

5.0

114

3.8

6.3

22

1.2

3. CONCLUSIONS The MODIS sensor, with a spatial resolution of 1 Ion, is marginally adequate for the estimation of FRP and the related estimation of the rates of burning biomass and gas and aerosol emissions at the regional and global scale. Though MODIS may miss a significant portion of small fires in comparison to BIRD, it underestimates the cumulative FRP of the fire scenes in Siberia only by - 4 % . The reason is that in these scenes the major part of FRP (and consequently of the fire pollutant emissions) are produced by large fires that are reliably detectable by MODIS. In cases of fires with a relatively small front depth, what is typical for the bush fires in Australia, MODIS may significantly underestimate the FRP by up to - 5 0 % as compared to BIRD. Clearly a combination of high spatial resolution and wide swath width (which is essential for a high observation frequency) is the ideal solution for a sensor aimed at fire detection, monitoring and characterisation, but this is difficult to achieve in a single instrument. Therefore, a reasonable strategy for the development of a fully operational space-borne fire monitoring system is the combination of: Wide-swath whisk-broom moderate-resolution spectro-radiometers, such as MODIS, in order to provide systematic global observations with a high observation frequency, with High-spatial resolution push-broom imagers, like the main sensors of BIRD but possibly with an further improved spatial resolution of -100-200 m, for detailed monitoring of the regions where fires have already been reported. The IR Element whose definition is planned by the European Space Agency (ESA) within the context of the Space Component for Global Monitoring for Security and Environment (GMES) shall be considered as a prospective European high-spatial resolution fire detection and monitoring sensor. 4. REFERENCES [1] Giglio, L., J. Descloitres, C.O. Justice and Y.J. Kaufman (2003) An Enhanced Contextual Fire Detection Algorithm for MODIS, Remote Sens. Environm., 87, 273-282. [2] Briess, Κ., H. Jahn, Ε. Lorenz, D. Oertel, W. Skrbek and B. Zhukov (2003) Remote Sensing Potential of the Bi-spectral InfraRed Detection (BIRD) Satellite. Int. J. Remote Sensing, 24, 865-872. [3] Zhukov, B., Briess, K., Lorenz, E„ Oertel, D., & Skrbek, W. (2005) Detection and analysis of high-temperature events in the BIRD mission. Acta Astronautica, 56, 65-71. [4] Lorenz, Ε. and W. Skrbek (2001) Calibration of a bi-spectral infrared push-broom imager. Proc. SPIE, 4486, 90-103. [5] Skrbek, W. and E. Lorenz (1998) HSRS - An infrared sensor for hot spot detection. Proc. SPIE, 3437, 1 6 7 - 1 7 6 . [6] Zhukov, B„ Briess, K., Lorenz, E„ Oertel, D., & Skrbek, W. (2003) BIRD Detection and Analysis of High-temperature Events: First Results. Proc. SPIE, 4886, 160-171. [7] Dozier, J. (1981) A method for satellite identification of surface temperature fields of subpixel resolution. Remote Sens. Environm., Vol. 11, 221-229. [8] Kaufman, Y. J., Justice, C. O., Flynn. L. P. et al (1998) Potential global fire monitoring from EOS-MODIS. J. Geophys. Res., Vol. 103. no. D24, 32215-32238. [9] Wooster, Μ., B. Zhukov and D. Oertel (2003) Fire radiative energy release for quantitative study of biomass burning: derivation from the BIRD experimental satellite and comparison to MODIS fire products. Remote Sens. Environm., 86, 83-107.

27

Figure 1. Fragments of forest fire mages acquired on 16 July 2003 by MODIS (4:21 GMT) and BIRD (4:46 GMT) at Lake Baikal; the detected hot clusters are coded using their FRP and projected on the MIR hand; the arrows indicate apparent false alarms in the MODIS data

b)

iU•

BIRD _ . MODIS

en 2 5 1 a> ιη υ 20 ο

15 -

Ο a> -Q 10 ;

ι Ζl·

:

Κ Jil

0 '

0.001 0.01

0.1

1

Effective fire area, Ha

10

0.1

10 100 FRP, MW

1000

10000

Figure 2. Effective fire area and FRP distribution from three MODIS and BIRD image swaths acquired in Siberia

BIRD

MODIS

• »

+

y· Castelo Branco

Γ

' *

j

.

P

-

•

i

-fl

iL ^

I

10 km I

Ζ .>

1

1.

I

100

10000 MW

Figure 3. Fragments of forest fire images acquired on 4 August 2003 by MODIS (11:30 GMT) and BIRD (12:04 GMT) in Portugal: the detected hot clusters are coded using their FRP and projected on the NIR band

10

•Κ "

. ,ί*-,

·., ·;

« Λ

i

t·

1

10 km

ί

1

10

1

100 1000 MW

Figure 5. Fragments of bush fire linages acquired on 5 January 2002 by MODIS (0:22 GMT) and BIRD (0:08 GMT) in Australia; the detected hot clusters are coded using their FRP and projected on the NIR band

0.001 0.01 0.1 1 10 Effective fire area, Ha

10 100 F R P , MW

1000

10000

Figure 6. Effective fire area and FRP distribution from the MODIS and BIRD image swaths acquired in Australia

30

EARTH OBSERVATION RESULTS OF FORMOSAT-2 FROM JUNE 2004 TO FEBRUARY 2005 Jeng-Shing Chern (M2) National Space Organization (NSPO) 8F, 9 Prosperity 1st Road, Science Based Industrial Park, Hsinchu, Taiwan 30077

ABSTRACT FORMOSAT-2 (formerly called ROCSAT-2) is a small satellite of 746 kg mass for two remote sensing missions: earth observation and upward lightning observation. It had been launched successfully from Vandenberg, California on 20 May 2004. The mission orbit is sun-synchronous of 891 km altitude for exactly 14 revolutions per day. This characterizes FORMOSAT-2 the daily revisit capability. For earth observation, the payload is an advanced high resolution remote sensing instrument (RSI) with ground sampling distance (GSD) 2 m in panchromatic (PAN) band and 8 m in multi-spectral (MS) bands. For upward lightning observation, the payload is an imager of sprites and upper atmospheric lightning (ISUAL). The first earth observation image was taken on 4 June 2004. From mid-July to late August, FORMOSAT-2 successfully monitored the disasters occurred in Taiwan Island caused by typhoons Mindulle and Aere. Then started from 28 December 2004, FORMOSAT-2 continuously observed the disastrous areas in South Asia due to an extremely strong seaquake and tsunami. This observation activity lasted four weeks. In order to provide images for rescue purpose in such an urgent condition, the National Space Organization (NSPO) of Taiwan opened its web site for free download during the first week. As to ISUAL, the first observation of sprite was on 4 July 2004. Up to mid-March 2005, 89 red sprites, 831 elves and 83 halos had been captured.

1 INTRODUCTION The FORMOSAT-2 was called ROCSAT-2 formerly. It is the first remote sensing satellite (RSS) owned by Taiwan for earth and upward lightning observations. After being launched successfully from Vandenberg, California on 20 May 2004, the early-orbit operations and in-orbit commissioning phases follow. The purpose of early-orbit operations is to transfer the satellite from 723 km circular parking orbit with 99.1 deg inclination to 891 km circular sun-synchronous mission orbit. It takes 11 days (23 May to 2 June) to complete the orbit transfer. Then during the in-orbit commissioning phase, the satellite system was checked with its performances validated and verified. After the completion of in-orbit commissioning, the satellite was claimed fully operational and ready to serve its customers. (1-2). There are two payloads onboard FORMOSAT-2: the remote sensing instrument (RSI) as primary and the imager of sprites and upper atmospheric lightning (ISUAL) as secondary. On 4 June 2004, RSI took its first image. Then on 4 July 2004, ISUAL captured a red sprite event for the first time. In this paper, the observation results from June 2004 to mid-March 2005 are presented. Major observations of RSI are disasters caused by typhoon Mindulle, typhoon Aere, and South Asia seaquake and tsunami. For ISUAL, a total of 89 red sprites, 831 elves, and 83 halos had been recorded. A preliminary concept of their distributions is discussed. (3-4)

2 RSI OBSERVATIONS 2.1 First Image The first image of RSI was taken on 4 June 2004. As shown in Figure 1, it is the seashore area of Hsinchu City, Taiwan. NSPO is located at the eastern part outside the figure.

31

2.2 Daily Revisit Capability For mission orbit at 891 km altitude, the satellite revolves the earth exactly 14 times a day. This means that FORMOSAT-2 repeats its ground track every day. W e call it the daily revisit capability. The local time at descending node (LTDN) is designed to be 10:00 am and currently it is 9:55 am. Consequently FORMOSAT-2 passes over Taiwan between 9:40 am and 9:56 am every morning. This property has two major advantages. With this capability, the day-to-day change of the observed areas can be recorded and traced for a long time. Databases of all kinds of useful information can be established. Another advantage is that the effect of weather can be reduced to the minimum degree. It is due to the reason that RSI repeats to observe the same areas at the same time everyday. Image of the specified areas can be taken once the weather condition allows. Figure 2 shows the images of Taiwan's Kaoshiung harbor taken in three consecutive days. (1-3)

2.3 Typhoon Mindulle Disaster Typhoon Mindulle swept Taiwan and surrounding oceans between 30 June and 1 July 2004. Heavy rain was brought and caused Taiwan's worst floods in 25 years. At least 22 people were killed and 14 were missing. The most serious disaster occurred in Taichung County. In Figure 3 the images of Shueili Township before and after typhoon ravage are put together for comparison. The left bottom image was taken by SPOT-5 on 26 June (before typhoon) while the upper two images were taken by FORMOSAT-2 in near infrared (NIR), left top, and natural color, right top, on 5 July (after typhoon). It can be seen clearly the river is flooded with much larger width after typhoon ravage. (5-7)

2.4 Typhoon Aere Disaster Typhoon Aere passed through Taiwan on 24 and 25 August 2004. Land slides and mudflows occurred in the Wufeng Township of Hsinchu County. At least 15 houses were buried with 31 residents killed. Figure 4 shows the landslides and mudflows image observed by FORMOSAT-2 and Figure 5 presents a picture taken on ground. (5-7)

2.5 South Asia Seaquake and Tsunami Disasters FORMOSAT-2 started to observe the huge disaster caused by seaquake and tsunami occurred in the South Asia region in the morning of 26 December 2004. With its daylily revisit capability, FORMOSAT-2 made the observation for 4 weeks from 28 December 2004 to 24 January 2005 as shown in Figure 6. Most of the observations are focused in Indonesia, Sri Lanka, India, Thailand, and Maldives. Two images taken are shown in Figures 7 and 8 for reference. Many other images are available from NSPO. (5-7) During the first week, NSPO put all images on its web site for free download. It is very real-time and very helpful for rescue purpose. Under such an urgent situation, time is the most important factor for rescue. Correct information must be provided as real-time as possible. NSPO just did not have time to negotiate the price, terms, and conditions with users who need the images so eagerly.

2.6 Other Observations Other observations such as the sport facilities for 2004 Olympic Game inside Athens City are shown in Figure 9. This image was taken by FORMOSAT-2 on 21 August 2004. For domestic applications, NSPO already planned to make a complete digital image of Taiwan every season by using FORMOSAT-2 observation images. The first issue had been published last year and the second issue is under preparation currently. Besides, take the advantage of daily revisit capability, many databases such as potential pollution areas, potential landslide and mudslide areas, etc. are being established.

32

3 ISUAL OBSERVATIONS 3.1 First Image on Red Sprites On 4 July 2004, ISUAL observed a red sprite over central Africa. A s s h o w n in Figure 10, this is the first observation result and the first image taken by ISUAL.

(8)

3.2 Observation Statistics From 4 July 2004 to 19 March 2005, the observation statistics is 89 red sprites, 831 elves, and 83 halos. The distribution of these 1003 events over the world is depicted in Figure 11. It is seen that: 1) the density of red sprites is relative high above land, such as America, Africa, and Asia, where Africa is the most dense area; 2) the distribution of elves is pretty uniform except above Africa the density is lower; and 3) the distribution of halos is pretty much the s a m e as red sprites, with Asia as the most dense area. ( 9 )

4 CONCLUSIONS After successful launch from Vandenberg on 20 May 2004, FORMOSAT-2 took the first earth observation image on 4 June 2004. Major disasters observed include those caused by typhoon, seaquake, and tsunami. In order to maximize the application of the high performance and daily revisit capability of FORMOSAT-2, N S P O has teamed up with Universities and other governmental Institutes to form a group. A digital map of Taiwan has been issued and shall be updated every season. Databases of potential pollution (high or low) areas, potential dangerous landslide and mudslide areas, etc. shall be established. The Environmental Protection Administration of Taiwan has nicknamed FORMOSAT-2 as an "Environmental Satellite". Statistically, about 500 segments of RSI observation can be made and 150 upper atmospheric events can be captured by ISUAL every month.

5 REFERENCES 1. Wang, H.C., Lee, L.C., Ling, J., and Wu, A.M., "ROCSAT-2 Remote Sensing Mission," Paper IAF-00-B.1.09, 51st International Astronautical Congress, Rio de Janeiro, Brazil, October 2000. 2. Chern, J.S., Ling, J., and Chang, Y.S., "ROCSAT-2, a Small Satellite for Two Remote Sensing Missions," Paper IAC-02-IAA.11.2.05, 53 International Astronautical Congress, Houston, October 2002. 3. Wu, A.M., Shiau, W.T., and Chern, J.S., "Ground Track Control of a Daily Repetitive Orbit, " Paper IAC-04-IAF-A.7. 05, 53rd International Astronautical Congress, Vancouver, October 2004. 4. Chern, J.S., "In-orbit Performance Verification of FORMOSAT-2," IAA-B5- 0402, 5th IAA Symposium on Small Satellites for Earth Observation, Berlin, April 2005. 5. CSRSR Document, "Satellite Image Analysis on South Asia Earthquake — Final Report," NCU, Chungli, and NSPO, Hsinchu, Taiwan, 4 January 2005. 6. DPRC Document, "2004/12/26 South Asia Earthquake Assessment — Aceh, Indonesia," NCKU, Tainan, and NSPO, Hsinchu, Taiwan, 4 January 2005. 7. IADC Document, "Analysis South Asia Earthquake," NTNU, Taipei, and NSPO, Hsinchu, Taiwan, 4 January 2005. 8. NSPO Document, "Final Review Report for ISUAL, Imager of Sprites and Upper Atmospheric Lightning," No. 9017-w0a, NSPO, Hsinchu, Taiwan, 30 Sept 2004. 9. Chen, P.C., "Distribution of ISUAL Observed Events," E-mail, forwarded by Liu, T.Y., NSPO, Hsinchu, Taiwan, 21 March 2005.

33

• Airplane

Ship

Figure 1 First image taken by FORMOSAT-2, Hsinchu County

Figure 2 Demonstration of FORMOSAT-2's daily revisit capability, Kaoshiung Harbor, Taiwan (left, 10 July; right top, 11 July; right bottom, 12 July, 2004)

34

, Jfcfi : i l : AX

UM&) :

iitoit ) a »1 -Τ Λ 5 Β

AT :

üLteil· ) a « - e η 2Q s

(SPOT

m%)

Figure 3 Flood at Shueili of T a i c h u n g C o u n t y c a u s e d by t y p h o o n Mindulle, b e f o r e (left bottom, 2 6 J u n e , by S P O T - 5 ) a n d after (left top, NIR; right top, natural, 5 July, by F O R M O S A T - 2 )

p~1

/

j k J

\V1

F i g u r e 4 L a n d s l i d e s a n d m u d f l o w s c a u s e d by t y p h o o n A e r e , W u f e n g of H s i n c h u C o u n t y

35

Figure 5 Ground image of typhoon Aere ravaged village in Wufeng

80 Longitude (deg E)

Figure 6 FORMOSAT-2 observed areas from 28 Dec 2004 to 24 Jan 2005

36

100

Figure 7 Destroyed seaside township in Ronong of Thailand

Sunshine seashore destroyed by tsunami in Maavaafushi Island

Estimated tsunami direction

Figure 8 Destroyed seashore in Maavaafushi Island of Maldives

37

Figure 9 Sport facilities inside A t h e n s City d u r i n g 2 0 0 4 O l y m p i c G a m e Image of Event 2004 186/21:31:15

50 100 1 50 200 250 300 Figure 10 First red sprite i m a g e c a p t u r e d by I S U A L (altitude k m vs. w i d t h km)

350

REAL TIME WEATHER FORECASTING: THE EGPM MISSION CONCEPT BY CARLO GAVAZZI SPACE TEAM V. Lo Rizzo*, F. Ortix*, M. Jarrett**, P. D. Erickson***, D. H. Warne**** *Carlo Gavazzi Space, Via Gallarate 150 Milano (Italy) **Sula Systems Ltd, Old Crown House, Market Street, Wotton Under Edge (United Kingdom) ***MDA, 13800 Commerce Parkway, Richmond, BC (Canada) ****EMS Technologies Ltd., 21025 TransCanadienne, Ste-Anne-de-Bellevue, QC (Canada)

ABSTRACT The EGPM mission has been foreseen as the European contribution to the Global Precipitation Mission, and is specifically aimed at improving global weather prediction' especially over Europe and Canada. In the mission phase A, the Carlo Gavazzi Space Team has assessed the feasibility of the EGPM mission, within the defined cost, schedule and environment boundary conditions. The team has performed the main system trade-offs to establish the baseline concept for the mission and has established a preliminary design of the system including all the relevant segments. The EGPM system has been demonstrated to be compliant with all the goal mission requirements. This paper describes the major outcome and selections of the study.

FROM MISSION OBJECTIVES TO MISSION ANALYSIS Mission Objectives The main objectives of the EGPM mission can be summarised as follows: •

Improvement of rainfall estimation accuracy.

•

Improvement of the delegability of light rain and snowfall over land in the Northern European area and over Canada. Monitoring of the flash floods over the Mediterranean coastal area. Increase coverage capability over the Mediterranean area for improving the weather forecast. Improve forecast skill of (Numerical Weather Prediction) NWP models with precipitation measurements. Provide near real time and real time data acquisition and delivery.

• • • •

40

Main Mission Requirements The major observation requirements taken into account during the mission phase A regard the timelines between data acquisition and data delivery. These have been considered as major drivers for the system selection and for the mission design in general. In order to accomplish the two main mission objectives, improving NWP models and the monitoring capability over the Mediterranean area, the data product is subject to the following requirements: • •

The mission architecture shall be compatible with a delay of less than 3 hours from image acquisition to level lb data delivery. The mission architecture shall be compatible with a delay of less than 15 minutes from image acquisition to lb data delivery over the European Mediterranean coastal zone.

This has led to a careful selection of the orbit altitude and ground stations network in order to accomplish the aforementioned objectives. EGPM Orbit The selected nominal orbit for the EGPM mission is a circular sun-synchronous orbit with LSTDN of 14:30 and orbit altitude of 520 Km. In order to meet the requirement on the incidence angle of the radiometer reference channel ( at 36.5 GHz), the instrument boresight is 47.44° with reference to the satellite nadir. The altitude has been selected considering the revisit time capability over the major area of interest (Europe and Canada), weighting it versus the quantity of fuel to be embarked on board of the spacecraft for the orbit maintenance. For the orbit selected, the orbit maintenance strategy required implies the altitude control every month of the mission lifetime. The orbit LSTDN, imposed by requirement, has driven the spacecraft configuration and in particular the solar array concept and instruments accommodation in terms of "fieldof-view" compatibility. Table 1 summarises the main parameters of the EGPM nominal orbit. Parameter Altitude

Value 520 km

LSTDN

14:30

Inclination

97.48°

Period

5702 sec

Repetition period (A + B/C) Table 1: EGPM orbit main

15 + 383/2500

parameter

41

latitude revisit time 520 Km

100 90

1/

80

70 »

60

g 50 40

"XL

J

30 20

\

\

.

10

0 34

39

44

49

54

59

64

69

deg

Figure 1: Spacecraft revisit time versus latitude

Ground Stations and Data Dump Scenario Two ground stations networks have been envisaged for the EGPM mission, in order to meet the different timelines requirements over different areas. The real time data delivery capability required by the system is accomplished making use of: • •

Matera ground station Villafranca ground station

For fulfilling the critical requirement on the delivery of near real time data product, the network selection has been driven by the operations availability and by the maximization of existing ESA infrastructures. Thus, the polar network is composed by: • •

Kiruna ground station Svalbard ground station.

The orbit altitude selected is compatible with these network configuration, as the revisit time over one of the two polar ground stations is always minor than the orbital period, ensuring the delivery of data with a maximum delay of 3 hours Two different strategies have been envisaged for the data product download, with reference to the two different timeline requirements. For the near real time data download to the polar ground stations network, the data acquired during one orbit will be stored on the solid state recorder and afterwards downloaded to the ground station in accordance with uploaded contact tables. For the real time data acquired over the Mediterranean coastal zone, the real time procedure is activated by longitude/latitude tables. When the spacecraft passes the boundary of the area of interest, a copy of the data stored on board will be downloaded in a burst.

42

Figure 2 European coastal zone on the Mediterranean sea area of interest. The MSCE will be in charge to define a set of longitude and latitude that corresponds to the start and to the end of the area of interest.

EGPM SPACECRAFT The EGPM Spacecraft is conceived to be able to operate, with generation of mission products, in an autonomous mode (without ground contact) for a period of 72 hours. To support such capability the satellite provides the possibility to store TC onboard (time tagged command) in order to execute them at a defined time. Moreover it provides a dedicated Contact Table to schedule the download periods of the P/L Data using the XBand link. The telemetry, self checking parameters and execution of procedures, will be stored on board and downloaded each day to the Kiruna ground station (during the maximum contact period pass over the ground station), by means of the S-band link. The on board mass memory will be able to store the data generated by the P/L for at least 3 hours and the housekeeping data for 24 hours. The main autonomous part of the EGPM mission will be the (Failure Detection Isolation and Recovery) FDIR. The implementation of autonomous FDIR gives the possibility to the satellite to "fail operational", without interrupting its tasks, if minor failures occur in the system. The autonomous FDIR will be structured in a hierarchical way with 4 independent levels described below from the lowest to the highest: •

• •

Unit Level FDIR. Each unit will be in charge of its own failure management through self checks (hardware checks, Out Of Limit (OOL) implemented and up loadable from ground). Subsystem level FDIR. In charge of managing a major failure at subsystem level. System Level FDIR. In charge of managing a failure at system level through the use of (On Board Control Processor) OBCPs.

43

•

System Alarm. If the OBCPs cannot cope with the system failure which has occurred, the system will enter in safe mode waiting for ground intervention and its instructions.

To meet the mission objectives the EGPM system makes use of a passive instrument (Microwave Radiometer) and an active Ka band instrument (Rain Radar). Microwave Radiometer The Microwave Radiometer (MWR) has double viewing capabilities with a swath aperture of 120° in the ram direction and 40° in the wake direction. Twenty one channels have been implemented (13 threshold channels plus 8 sounding channels) in a frequency range between 18.7 GHz and 150 GHz. The selection allows the achievement of specific performance requirements of the mission, such as the measurement of light rain and snow in the high latitude region of Europe and Canada. The MWR design has been derived building on the experience gained from previous programmes, including MIMR. The proposed instrument meets the users' requirements with low risk and minimum cost. The design has been tailored for a small spacecraft application and for compatibility with the Precipitation Radar.

Figure 3 MWR configuration

The MWR comprises a rotating unit (rotor) with other units accommodated within the spacecraft. The instrument's primary reflector is fixed in its orbital configuration during launch. Launch protection is provided by the four (Hold Down and Release Mechanisms) HRDMs and by local offload devices within the scan mechanism. This

44

approach offers significant advantages over a deployed antenna including lower risk and cost. Key features of the instrument are: • • • • •

•

• •

Scan rate ~31.5 rpm providing contiguous footprints along track at 36.5GHz Integration times selected to provide "circularised" ((Effective Field Of View) EFOV No compensation of exported disturbances provided by instrument. Antenna aperture is 1,2m for 520km orbit Shared feedhorns for 18.7GHz and 23.8 GHz, and for 52.5 GHz and 118.75., All other feeds are single frequency. Two sets of horns and receivers are provided at 89 GHz and 150 GHz to improve spatial sampling. Direct detection channels for 18.7GHz and 23.8 GHz. All others are heterodyne (36.5 GHz is heterodyne to allow inclusion of additional filtering at IF for protection from radar interference). Scan mechanism uses DC brushless motor with dry lubrication bearings and Inductosyn resolver. Power and Signal Transfer uses rolling elements for power and capacitive coupling for signals.

Rain Radar The Rain Radar concept selected is fully compliant to system requirements. A short pulse architecture has been selected in order to implement a low risk solution without impacting overall performance. One of the major issues related to the radar has been the assessment of its compatibility with the radiometer. A concept has been assessed during the study and its feasibility demonstrated by analysis supported by limited testing. Nevertheless, the instrument is compatible with the spacecraft architecture, in terms of accommodation and field of view analysis. Considering the state of the art, the rain radar baseline presents low risk and complexity, as no brand new technologies are required to implement the architecture selected and to meet all the mission requirements. In general, most of the radar elements are relatively standard in concept and implementation technology. Functionally the instrument has been divided into three main subsystems; the antenna, the RF subsystem and the digital subsystem. The design will support generation of three radar beams operating at 35.505 GHz with Pulse Repetition Frequency (PRF) optimally selected for the operational altitude of the spacecraft. The basic design choices were between the long and short pulse radar implementations. The main differences between the design choices are in the transmitter power and transmit/receive signal processing. A long pulse option requires a digital waveform generator. Subsequently the digitally coded transmit pulse must be converted to an analogue signal at the radar transmit frequency and amplified by a low power Solid State Power Amplifier (SSPA) to a

45

transmit power level of approximately 2 Watts. This method requires matched pulse compression of the received signal. The short pulse option can be implemented using well-matured High Power Amplifiers technologies (TWTA or Klystron). Since the short pulse implementation does not require receive pulse compression there is no need for digital transmit waveform generation and the high power amplifier can therefore be directly controlled to modulate the CW signal at the required pulse width and repetition frequency. The receive signal processing that has to be performed on the detected signal is limited to sample averaging of the received signal power level. The selected detection approach uses linear diode signal detection followed by a 12-bit Analog to Digital Converter (ADC). Using this detection method the signal voltage to power (squaring) will be performed after digitization.

EGPM Platform The EGPM platform is fully flexible and capable of implementing different payload architectures, i.e. with or without rain radar, without critical impact on the overall design. The design inherits the (Carlo Gavazzi Space) CGS experience from past and running programs, implementing a low risk and cost effective solution. The EGPM platform is composed by two main subassemblies, the bus and the solar array. The solar array is articulated around its longitudinal axis. The Microwave Radiometer is composed of a rotating part, including the reflector, the feed horns, the balance mechanism and the receivers, and a fixed part, hosted in the Bus structure, comprising mainly the Power and data handling unit (PDHU), scan mechanism and hot and cold calibration sources. The rotating part is accommodated on the Zenith face of the Bus allowing the required field of view both in ram and wake directions. The rotating part of the radiometer is mounted on the external side of a dedicated Aluminium alloy machined panel. The Rain Radar is on the Port side, hinged on the top and fixed during launch also on the satellite's bottom side, close to its Feed horns, which are not deployable. After launch, the Radar reflector deploys facing almost nadir and the required nadir pointing direction is satisfied. Like the radiometer, the Rain Radar electronic units are accommodated inside the Bus. The Radar can be considered integrated with the Platform with the following advantages:

46

• •

The envelope is optimised since the units are accommodated exploiting the available volume inside the Bus. The Reflector and the Feed are fixed to the same panel, that is also a structural part of the Bus. Thus thermo-mechanical stability properties are optimised both for the Antenna and for the Platform.

The satellite power is supplied by the solar array that is wrapped on the satellite's starboard side, and fixed by non-explosive release fasteners on the corners during launch. The array is composed of 4 panels hinged together, folded and stacked during launch. Once deployed, the solar panel is hinged by the one-axis solar array driving mechanism. The driver guarantees sun-facing tacking during the spacecraft orbiting, rotating around the axis at 51 deg with respect to the satellite nadir vector. A rechargeable battery is used to supply power to the satellite during the eclipse periods and during LEOP operations. The overall spacecraft design has strongly addressed the requirement for single point failure avoidance. Thus, this has driven the definition of the redundancies in the system. Nevertheless, the EGPM spacecraft has a few single point failures that have been defined for equipments whose redundancy is not feasible for implementation (e.g. momentum wheel). The platform avionics is a highly integrated unit. A single unit, the PDMU (Power and Data Management Unit) contains the power conditioning and distribution electronics, and the data handling electronics. The platform contains two identical units for redundancy. All the other equipments of the satellite are interfaced to the PDMU. The PDMU offers a set of standard interfaces and the payload instruments are connected to it by means of a MIL-1553 bus or RS-422, in order to reuse some existing instrument hardware development. The attitude control is performed by a set of reaction wheels supported by magnetic torquers. The attitude determination is made mainly by two star sensors and laser gyros for the operational phase, magnetometers and sun sensors for the course pointing and safe mode. A dedicated momentum wheel is employed to compensate the angular momentum about the nadir axis due to radiometer rotation. The orbit maintenance control is performed by Hydrazine thrusters. The Propellant tank (a 40kg Hydrazine gas pressurized tank) is fixed to the middle fuselage frame, in order to be close to the satellite COG, thus reducing oscillation disturbances. The two set of thrusters located on the wake side and on the ram side, are at a reasonable distance from the tank reducing pipes mass, thennal and vibration flux disturbances and resistance A GPS receiver is included in the AOCS architecture. It provides precise orbital position to help the orbit control and the ground contact management. The GPS provides also synchronisation of the on board time with UTC. The communication subsystem is composed of an S band transceiver for satellite control and an X band transmitter for payload data download. The X-band antenna, and the nominal S-band antennas are placed on deployable booms, pointing toward nadir

47

direction and have a 124 degrees open FOV. In order to guarantee the possibility of achieving ground station contact in a non nominal attitude (e.g. during the LEOP or during safe mode), two additional S band antennas are located on Bus structure pointing toward the nominal wake and ram directions. The platform is a skin-frame structure composed by aluminium alloy panels, corner posts and a core reticular frame in the centre. The corner posts work as joint elements on the four corners of the Bus, fixing the lateral panels and the Bottom and Top panels. This coque is stiffened by a horizontal honeycomb panel and a core reticular frame inside the Bus. The middle panel gives the opportunity as well to divide the Bus external closing panels in two panels per side. The solution maximizes the volume for the accommodation of the internal units and accessibility. The EGPM platform thermal control will be of passive type, mainly based on rejection to space of heat generated by units. The main spacecraft budgets are summarised in Table 2 Element / Parameter

Description

Mass (Dry)

618 Kg

Power

492 W

Attitude

Nadir pointing, 3-axis stabilization by reaction wheels

Pointing accuracy

Pointing accuracy (instrument boresight): < 3.4 mrad

Pointing stability

Pointing stability > 0.5 mrad

Pointing knowledge

Pointing knowledge (instrument boresight): < 2.4 mrad

Propulsion

Mono-propellant propulsion type 40 kg propellant for orbit keeping (5 year life)

Table 2 EGPM Spacecraft

budgets

EGPM GROUND SEGMENT The main goal of the EGPM Ground Segment is to receive data from the EGPM Space Segment and to produce and deliver atmospheric products of precipitation rates to End Users in a timely and cost effective manner. The EGPM Ground Segment (EGS) consists of a Command and Data Acquisition Element (CDAE), the Mission operations and Satellite Control Element (MSCE) and the Processing and Archiving Element (PAE). Their main tasks are the tracking and communicating with the satellite,

48

commanding and monitoring of the spacecraft status and the acquisition, and processing, archiving and delivery of the science payload data to End Users. The EGPM ground segment will also implement an effective interface with a Science Data Centre, external to the segment, and an interface with the product processing system of the global GPM system. The PAE functions generate level 0 and lb products for delivery to End Users and Science Data Centres, where level 2/3 products are generated. End Users receive products from the PAE automatically via electronic file transfer to meet their real-time and near real-time operational needs for precipitation data. End Users may also browse the catalogue and place orders for products already archived or requiring reprocessing. Support data sources of External Calibration Support data will provide specific field data occasionally to assist in the calibration of the instruments. The EGPM mission is also a contributor of data to the overall GPM mission and therefore must provide calibrated sensor data within 3 hours to the GPM Precipitation Processing System. EGPM LAUNCHER

The Phase A study has identified several launcher candidates for the EGPM spacecraft injection in orbit. Considering the system trade-off, two launchers have been selected as baseline: • •

COSMOS PSLV

The baseline launcher will be defined, in accordance with the Agency indications, during the next steps of the EGPM program.

49

PROGRESS IN SMALL SATELLITE TECHNOLOGY FOR EARTH OBSERVATION MISSIONS Alex da Silva Curiel, Andrew Cawthorne, Martin Sweeting Surrey Satellite Technology, Ltd Surrey Space Centre, Guildford, Surrey, GU2 7XH, UK http://www.sstl.co.uk [email protected]

ABSTRACT The value of Small Earth Observation satellites can be measured directly by considering the lifecycle cost of the system, and the amount of valuable data delivered in its lifetime. In practice this needs to be validated by a science case or business plan, which assigns a market value to the data, and considers many of the underlying issues. These include the applications of the data, its quality, whether the data is perishable or retains its value, and whether there are alternative means of obtaining the data. Increasingly, small Earth Observation satellites operating within groups are considered to offer good value and utility, and are able to offer services that cannot be practically offered by deploying larger spacecraft. This is as a direct consequence of the growth in capability of small satellites in recent years. A constellation of small satellites has the benefit of increasing temporal resolution, and when coupled with significantly lower unit costs, can support niche markets that are otherwise economically unviable. It is anticipated that small satellites will make a significant input into the economic viability of commercial Earth Observation. Against a backdrop of benchmarking technology and performance trends in small Earth Observation satellites, some of the advanced small satellite platform designs being developed at Surrey are described. It is demonstrated how such spacecraft can be used in high temporal resolution missions, that can be dynamically adopted in orbit to meet both peace and crisis time applications. TRENDS IN SMALL EO MISSIONS There has been significant debate in the small satellite community regarding the relationship

50

between small satellites' and larger satellites. Are small satellites a disruptive technology, and will they replace larger satellites eventually? Or are they complementary, and will be used to address applications not viable with larger spacecraft? There is unlikely to be an unambiguous answer to this, as it is highly dependent on the application and market factors. No doubt miniaturization will permit some space missions eventually to be accomplished with much smaller spacecraft than today, and consequently the cost-of-entry to execute space missions may reduce. Figure 1 provides some evidence that smaller spacecraft generally cost less then their larger counterparts. This plot uses spacecraft and programme costs as published, normalised to US$ (2003). The necessary physical apertures will still drive the size of some Earth Observation missions. Although eventually synthetic aperture generation may become possible by using spacecraft in formation, at the moment these missions are best accomplished with a single spacecraft. Certainly in some cases, the overheads involved in deploying and operating a satellite may dictate that larger spacecraft provides a better economical solution if data return, or communication bandwidth are of primary concern. This is exemplified by the fact that most Geostationary communications satellites have grown to be quite large in a commercial environment. As the value of a mission is largely dictated by the performance for a certain cost, capability of these small satellites must also be examined. If one examines the trends in capability of small Earth Observation satellites, it becomes apparent that this has advanced rapidly in recent years from early technology demonstrators, towards a point where 1 For the purpose of this paper small satellites are defined here as those that have a wet mass less than 500kg.

Spacecraft mass and Programme cost

10

100

1000

Total Mass (kg) Figure

1. Cost of spacecraft

and complete mission plotted

against spacecraft

mass

missions increasingly provide operational services. Examples include the DLR BIRD satellite, the ESA PROBA satellite, and SSTL's Disaster Monitoring Constellation (DMC). In Figure 2 it is demonstrated that alongside the steady progression and improvement in the Ground Sampling Distance for civil optical imaging missions, small satellites are catching up rapidly in this capability. This indicates that small satellite capabilities in instrument accommodation, miniaturisation of avionics, data return, power, and attitude control are all improving. As a result small satellites are starting to address operational applications traditionally carried out with larger satellites, and are encroaching on the larger markets for high-resolution imagery. Spacecraft included in the plot for spacecraft weighing under 500kg include DMC, Orbview-3, BIRD, UoSat-5 & 12, KITSat-1, 2 &3, Posat-1, Fasat-Bravo, Thai Phutt-1, Tiungsat, Tsinghua-1, Earlybird, Saudisat-2, as well as several spacecraft under contract and due for launch over the next year, namely TOPSAT, DMC+4, Rapideye, and RazakSAT.

Figure 2. Improvement

of Ground Sampling

(a) All Satellites (b) Small

Distance

Satellites

51

H O W W I L L S M A L L EO S P A C E C R A F T MISSIONS P R O G R E S S ? Key demands for improvement in Earth Observation missions includes better resolution, greater spectral coverage, and faster data return. For commercial missions there is a need for improved costeffectiveness, and in scientific missions the data quality is often the main driver. Although small satellites can effectively address some of the niche applications, one of the areas where it is likely that small satellites will flourish is where there is a strong requirement and benefit in using groups of satellites; constellations, formations and swarms. These become particularly attractive where launcher capacity can be used efficiently, for instance for deploying an entire plane of spacecraft simultaneously. It is often prohibitively expensive to deploy multiple large Earth Observation spacecraft for such applications. Groups of spacecraft offer unique capabilities that are often difficult or impossible to achieve through alternative means, for instance the requirement to enhance temporal coverage. Individual spacecraft can improve their temporal resolution through off-pointing the instrument. This has an effect on the achievable Ground Sampling Distance at the extremities of the Field-Of-Regard (FOR) for such repeat visits, as highlighted in Figure 3. The effect on the effective instrument swath width is plotted in Figure 4 (assuming a fixed instrument field of view). In this case an instrument with a nominal G S D of 2.5m is assumed, leading to degradation in the GSD when pointing further offtrack (but dependent on the terrain slope). Due to the resulting distortion, there is a fundamental limit to h o w far the instrument can be off-pointed to serve the target application. This off-pointing requirement can be relaxed by increasing the number of satellites, and it is readily shown that in practical cases there is a virtually linear relationship between the temporal resolution and number of satellites deployed. Hence the use of constellations of imaging satellites does not only improve the revisit time, but also some aspects of the quality of such high temporal resolution data.

o f f - p o i n t i n g effect o n G S D

\ \ ^

-50

J

8.00 6.00 -00

Ν.

-40

1

-30

-20

-10

0

10

20

30

40

50

o f f - p o i n t i n g (Deg)

Figure 3. Ground Sampling distance off-pointing effect on swath width

100.0 90.0 80.0 70.0

-50

-40

-30

-20

- • - 4 5 0 km

-10 0 10 20 off-pointing (Deg) •

567 km

686 km

30

40

50

800 km

Figure 4 Off-pointing effect on the field of regard Due to aperture constraints on small spacecraft, higher resolution imagers will typically have a smaller window of local time in which imaging can be achieved, for example, between 10-2pm. Beyond these times of day, the image quality may substantially reduce. It can be observed in Figure 5, when off-pointing by 50 deg at an altitude of 567 k m the local time will change f r o m 11am to 10:20am, hence larger off-pointing angles may also lead to sun

illumination angles that can not be supported. Of course the level and nature of image quality required depends on the application hence, strict limits cannot be discussed here.

Other characteristics that can be well addressed by constellations of small low-cost satellites are freshness of the data, synopticity (the amount of data captured in one instance) and affordability of the data products. By examining the broad applications of spatial and temporal resolution (Figure 7), it becomes clear that small satellite capability has now matured to the point where a wide range of missions can be addressed. The capabilities of small spacecraft in constellation are already being exploited by emerging missions such as the operational D M C mission and planned RapidEye constellation. These missions demonstrate that new markets and business can be created.

Figure 5 Local time shift Daily coverage for any (illuminated) point on the globe with suitable spatial resolution can be achieved with four spacecraft dispersed in a 600km orbit. This is depicted in Figure 6, where each colour track represents the coverage of a single spacecraft. A single spacecraft with the same Field-Of-Regard can only achieve daily revisit on a regional basis, or global coverage with lower temporal resolution.

φ

t>

i

WeaLMer

β

•»••Ss Assessment & Planhing

Figure 7. Applications map - Temporal and spatial resolution TECHNOLOGY TRENDS AND NEEDS

Figure 6. (a) One day coverage provided by four spacecraft at 600 km with 30 off-pointing capability compared with (b) single satellite capability.

State-of-the-art small satellites are currently being developed and deployed with G S D of 2- 5m, and technology developments will enable small satellites to still improve turther. This permits small satellites to be employed across the broad range of traditional applications in optical Earth Observation, but their lower cost makes niche applications economically viable. In examining the applications for higher

53

temporal resolution, it is clear that small low cost satellites in constellation can also enable new applications, competing with aerial surveys.

carry GPS receivers and star trackers. There is also still a need for good Digital Elevation Models (DEM).

It is now worth considering some of the other issues in the move towards higher resolution small satellites. The most immediate trend in small EO satellites has been the improvement o f Attitude Control and Determination Systems. Higher pixel line rates and better G S D lead to a greater need for better spacecraft stability. This is readily achieved with small spacecraft, which have traditionally used simple structures and few deployables, yet this is threatened by other trends in some small satellites with the move towards deployable solar panels and liquid propellants. Propulsion systems have not traditionally been flown on small satellites, but recent moves towards the use o f constellations such as the S S T L Disaster Monitoring Constellation have introduced such capability for station keeping. For higher resolution missions, contamination issues and sloshing and becomes a concern, yet more capable systems will eventually enable more advanced applications of constellations, for instance by deployment of spacecraft from a single launch vehicle into multiple planes. Small spacecraft are also well suited to missions requiring a high degree o f agility. In most cases the whole spacecraft is maneuvered allowing the instrument to be rigidly mounted to the platform. B I L S A T - 1 carries Control Moment Gyros to achieve rapid off-pointing, D L R - T U B S A T has demonstrated interactive target tracking, and the TOPSAT mission employs Time Delay Integration maneuvers to dwell for longer over a selected target. The accuracy o f the geo-location is also critical in some missions. Indeed, the swath width of a high resolution-imaging camera (GSD . = - 9,3 μιη)

Fig. 1 The BIRD Spacecraft with Payload The infrared diode arrays were made of CdxHgi_xTe. This material has the great advantage that the desired wavelength region of the detectors can be tuned changing the mixing ratio x. In the consequence all components of both IR cameras are fully identical. This reduces remarkably the efforts of the manufacturing and testing of the infrared cameras.

104

All imaging sensors are designed as push broom cameras. This is the only way to implement a veiy compact, robust and at the same time a high-performance payload system compatible to limited recourses of a micro satellite. For the BIRD mission a new approach of an on-board data processing system was realised. The main idea of this on board data processing system is to reduce the downlink data stream by generation of thematic maps. This reduction can be achieved by a multi-spectral classification of the available sensor signals on board of BIRD. A neural network, implemented in the neuro-chip „Recognition Accelerator Nil000" from Nestor Inc, executes the classification process. WAOSS-B fB/äväength Focal length II k;i(i))I;vip'v",»·· f-number Detector ' DetectorvEOöling Piwl skr Pixel number : Quantization Ground pixetsize;1 GSD i Swath width

Vis- 600-670 nm NIR: 840-300 nm

;21.65 mm ISffiK 2.8

„CCD linelaH päSSivggo ?#:f: f μφ χ f pm 1ϊ

1 is m 185 m 533 km

Two-channel infrared sensor system MIR' 3.4-4.2 pm p R : 8,5-9 3 um 46.39 mm 1®1;· =

ΐΐο =(CdHg1g Arrays ·;Κ '^Stirling, 80-100 Κ

Κ0 pm xfO pm 2vx;51::2 staggered 14 bit (for each exposure)^

Wt m 185 m 190 km

Tab. 1 Instrument Characteristics of the BIRD Payload

4. ORBIT EXPIRIENCE AND RESULTS OF TECHNOLOGY EXPERIMENTS According to the need to decrease the spacecraft costs and keep at the same time high performance characteristics of the spacecraft bus the BIRD satellite demonstrates new developed technologies at moderate costs in space. 4.1

Low-Cost Star Sensor

An autonomous star sensor is necessary for the high precision attitude information. In the field of micro-satellite technologies there was a need for small and autonomous working star sensors in a low-cost price range. The BIRD star sensor development in close co-operation between Jena-Optronik GmbH and DLR filled this gap in the market. The sensor (see Fig. 2) has a robust and compact electro-optical design with a total mass of 1.2 kg. The sensor consists of a CCD matrix camera in combination with an internal star catalogue and image analysis software for star identification and search.

105

Fig. 2 Star Sensor Astro 15 The test in orbit showed that the 2 star cameras determine their orientation reliably and with the required precision. Their data are used as well for the on board attitude control, as well for the geo- referencing of the BIRD images to deliver geo coded data products. 4.2

High Precision Reaction Wheel

For BIRD a reaction wheel system with 4 wheels in combination with 2 x 3 magnetic coils are applied. Up to now, there does not exist a big choice of different types of reaction wheels on the market. Based on the reaction wheel development of the Technical University of Berlin (Prof. Renner) a new type of high-performance reaction wheel for micro-satellites was developed for BIRD in close co-operation between Astro- und Feinwerktechnik Adlershof GmbH, TU Berlin and DLR. These reaction wheels (Fig. 3) are characterized by • high control precision by means of smart control electronics, • low level emitted vibrations, due to the advanced mechanical design and the high level of alignments and balancing, • integrated electronics and robustness of assembly.

Fig. 3 BIRD reaction wheel assembly

106

The reaction wheels were space-qualified and intensively tested including long-term behaviour in labs, test facilities and test chambers. But the high precision of control activities in space cannot be verified under lab conditions. For this the experimental determination of control activities in space was necessary. The active space application could verify the required specifications. 4.3

Spacecraft Bus Computer

The Spacecraft Bus Computer (SBC, see Fig. 4) controls all activities of the satellite bus. The SBC receives stores and processes the commands, gathers and evaluates the housekeeping data of all subsystems and partially of the payload and controls the telemetry and science data formatting and transmission activities on-board. Furthermore the Spacecraft Bus Computer is also the attitude control computer of BIRD, this means the complete hard- and software of the attitude control system is embedded in the SBC. In summary the SBC of BIRD has to fulfil the following general requirements: • handling of a lot of different electrical and data interfaces • controlling of different software tasks and algorithms • controlling of time depending activities and software loops • high processing performance (near real time) • processing of high and low data rates • monitoring of all spacecraft activities including the SBC-activities • failure tolerance • high robustness of all monitoring and control activities. • i s ; · 5» ] Ι ί Λ fefttsM Wh η Ü m ν··. P · « ui us ί

if ' ' I s !SHgr I

SM

•

ymi f

•Iii S Ίΐ *,··?_, :H I

Λ .·»!· •Wf'vy^Jfe ι s Ι Β Ε Ι ' ; ; " ' "

Fig. 4 BIRD board computer by Fraunhofer FIRST Finally the SBC is a computer system characterised by a high complexity resulting from its different interfaces to all subsystems, its components and its software modules. The SBC has to work in a very reliable, robust and failure tolerant way. However, the property of high complexity stands in opposition to the requirement of high reliability. To solve this conflict two measures were taken: • implementation of measures to improve the reliability and robustness of the SBC Board (latch-up detection and protection, redundant processing and memory structures) • implementation of a highly redundant structure of 4 SBC boards and watchdog circuits for failure detection and recovery. The new hardware technologies of the Spacecraft Bus Computer had to be investigated with reference to their behaviour in the real space environment.

107

The board computer hardware was working from the very beginning without problems. The same is true for the implemented Unix-based operational system. Some bursts of high-energy radiation and extremely high particle density caused a switch to the redundancy structure or into the safe mode. These events have demonstrated the functioning of the redundancy structure as well the latch-up safety control implemented on the computer board. 4.4

The Infrared Cameras

With respect to the limited project resources the most challenging problem of the payload design was the accommodation of two machine cooled infrared push broom cameras. At first it was necessary to find suitable detector arrays, which fit the financial budget of the mission. The mission science program requires long linear arrays, capable to operate in a push broom mode avoiding any mechanical scanner, which disagree with the philosophy of a small satellite. Long linear arrays will be very seldom used in commercial infrared cameras and, therefore, it is difficult to find an adequate off the shelf product. The development of a new chip was not within the financial budget. Fortunately the arrays were found and offered by GEC Marconi (today BAE Systems) as a off the shelf product. It is worth mentioning, that the situation has not changed up to now. In the FUEGO-2 program a linear array was required, twice longer than the BIRD detector. Again for financial squeezes such an array could not be realised, whereas different European companies showed the feasibility of such an array. Meanwhile the detector technology offers the opportunity, to realise even two or more spectral bands on one chip. With such a chip the efforts for the IR pay load would be only a half of the BIRD IR pay load reducing the efforts for the camera equipment. While the chips of the BIRD IR instruments are off the shelf, the encapsulation and the integrating of the cooling engines required a special development to be compatible with the satellite recourses. The related commercial solution was voluminous, heavy and power consuming. The detector provider proposed to use small cooler devices, used in the CLEMENTINE mission. The problem of such small cooling devices is mainly their short lifetime. The expected lifetime of the used coolers was 8000h and this figure was compatible with the originally designed mission duration. Further more, all components of the payload segment are tuned for a 10 minutes measurement cycle per orbit what's fully sufficient to cover a selectable, desired area of interest. In the previous 1000 days of the BIRD mission the cooling engines have shown no degradation. In Fig. 5 there is demonstrated the compactness of the realised IR sensor head including the optics.

108

Up to now the detector arrays are working excellent without any remarkably degrada-

5. CONCLUSIONS The BIRD spacecraft bus is a new adaptive platform suitable for further science or Earth watch missions supplementing already planned missions. The applied design to cost philosophy could be justified. Considering the newest technological development a further cost optimisation is possible. Whereas new developed products may have a higher price than the older one, due to their improved properties they can reduce the material and personal efforts in manufacturing and testing of new devices and reduce the costs in that way.

6. REFERENCES [1] James R. Wertz and Wiley J. Larson (editors) Reducing Space Mission cost Space Technology Library Kluwer Academic Publishers Dorndrecht/Boston/London 1996 [2] Rainer Sandau (Principal Editor) 1AA Position Paper on Cost Effective Earth Observation Mission Prepared by the Study Group on Cost Effective Earth Observation Missions 2002-2004 [3] E. Lorenz, W. Bärwald, Κ. Briess, Η. K a y a l , Μ. Schneller, Η. Wüsten Resumes of the BIRD Mission Proceedings of the ESA 4S Symposium: Small Satellites, Systems and Services 20-24 September 2004 La Rochelle, France [4] FIRES, Fire Recognition System for Small Satellites, Phase A Study, DLR Institute of Space Sensor Technology, Berlin and OHB-System, Bremen, Germany, 1994

109

Demonstration of a Semi-operational Fire Recognition Service Using BIRD Micro-satellite - DEMOBIRD Jesus Gonzalo, Gonzalo Martin de Mercado Ingenieria y Servicios Aeroespaciales, S.A. (INSA), Spain Paseo del Pintor Rosales 34, 28017 Madrid, Spain Eckehard Lorenz, Dieter Oertel German Aerospace Center (DLR), Germany Jose Luis Casanova Laboratorio de Teledeteccion de la Universidad de Valladolid CI Real de Burgos s/n, Valladolid, Spain Miguel Aguirre, Wolfgang Leibrandt, Gerhard Billig European Space Agency ESTEC/The Netherlands, ESOC/Germany 4-

Abstract

FUEGOSAT is the European Space Agency initiative to improve the management of forest fires by means of space remote sensing. The programme is executed through different actions, one of which consists of the development of a semi-operational end-to-end fire recognition service. The Bi-spectral IR Detection (BIRD) satellite, a technological demonstrator designed for the recognition of hot spots, is an excellent precursor for prospective data sources to fill the gap of existing Earth observation satellites and future dedicated satellites that will provide the definitive tool for fire monitoring. Representative regions were selected for a first demonstration in South Europe. After performing the exercise during 2003 summer campaign in Spain, BIRD satellite sensors have proved to be excellent for fire detection and monitoring, and good enough for burnt area data generation, as fire management operators have confirmed. It was possible to generate in time a whole catalogue of products valuable for the decision makers based on BIRD real time transmission data. Special mention has to be done to the hot spot inventory product, able to detect very small fire outbreaks in a reliable way. A

Introduction

to the FUEGOSA

Τ initiative

at ESA

The FUEGOSAT framework is the ESA initiative to pave the way for the development of future satellite-based operational services for the management of emergencies. Under this initiative, the DEMOBIRD project was proposed to develop an end-to-end demonstration of the use of satellite remote sensing applied to the forest fire problem in all its phases, especially detection and monitoring. To achieve this objectives the Bi-spectral InfraRed Detection (BIRD) satellite, a technological demonstrator of the German Aerospace Centre (DLR), designed for the recognition of hot spots by means of high resolution infrared cameras, was used. It proved to be an excellent precursor of future dedicated systems. This paper explains the first ESA supervised end-to-end demonstration exercise of semioperational space borne fire recognition in Mediterranean countries using a dedicated fire sensor system on a small satellite.

110

4

Introduction

to the

Demonstration

The forest fire phenomena in temperate latitudes represent a serious economical and ecological problem, especially in some areas as the Mediterranean ones in Europe. In many occasions, the current fire fighting systems are hardly good enough to manage this problem, as can be seen from the statistics of forest fires and area burnt every year. The total area burnt during the last decade is more than 4.5 million of hectares, an area larger than that of Denmark. Some of the major problems related to forest fires are the loss of ecosystems and the deforestation, erosion and desertification, which encompasses substantial economical losses which can be quantified at around 3000 EUR per burnt hectare only in Europe. The use of new technologies, such as space systems equipped with infrared sensors, can help in tackling this problem The key project objectives can be summarised as follows: •

Consolidation of the user requirements of an operational system aimed to provide support in forest fire monitoring using space Earth Observation data; the active participation of the end-user has been essential to establish the proper starting basis for this research activity; several reference user were selected (Xunta de Galicia, Spain) to operate the pilot demonstration system and drawn conclusions for future developments.

π

Development and construction of algorithms to provide a tool for the operational use of data from BIRD and other satellites, including:

n

-

Automatic data reception and image processing at the user facilities without human intervention

-

Generation of value-added products with minimisation of interactions with the system operator, but allowing control of processing parameters through the system Graphical User Interface

-

Automatic local database maintenance for a specific fire emergency scenario reducing as much as possible human intervention

-

Integration with existing GIS's, customised to the specific functionalities of the system

-

Product archive available for dissemination purposes

Execution of a pilot demonstration, including the following targets: -

Demonstration of the suitability and usefulness of the information derived from the combination of BIRD data with other satellite and GIS sources

-

Demonstration of system performance at user premises and, as a consequence, of the FUEGOSAT service potential capabilities

-

Simulation of a satellite data provision service, based on two processing centres, the first to acquire and pre-process the data and the last for dedicated fire product generation and data server for the final user tool.

-

System evaluation with active participation of reference users, recommendations and lessons learnt in view of future developments, such as FUEGOSAT

Following the analysis of operator requirements, the concept depicted in Figure 1 was developed. The design includes an Earth Observation data processing subsystem called External Data Gateway (EDG). The function of this EDG is to serve as a centralized EO data storage for the operator network to download products, regardless the source satellite. The project was kicked-off in January 2003 with the campaign planned from June to September 2003 in Galicia. Up to eight days along this period were selected to perform the campaigns. The concept proof was run the 30th of May in Galicia. Additionally, monitoring products were generated for fires in Portugal around 15th August to be presented to the users off-line.

111

i- Overview of BIRD system DLR's BIRD micro-satellite is a technology demonstrator of new infrared push-broom sensors dedicated to recognition and characterisation of thermal processes on the Earth surface. BIRD was piggy back launched with an Indian Polar Satellite Launch Vehicle (PSLV) on October 22, 2001 and it was successfully operated by DLR until February 2004, when the a malfunction in the BIRD attitude control system interrupted its continuous operation. The BIRD main sensor payload consists of: -

A new two-channel infrared Hot Spot Recognition System (HSRS) designed to avoid saturation, preserving at the same time a good radiometric resolution

-

A Wide-Angle Optoelectronic Stereo Scanner (WAOSS-B)

Due to its higher spatial resolution, BIRD can detect fires with the area a factor of seven smaller than AVHRR and MODIS radiometers. However, it has to be considered that BIRD'S re-visit time was about once every 10 days. Repeated observations of one area within 3 days was possible by satellite cross track tilting.

4- Data Communication Concept BIRD ground segment is limited to reception stations in Germany. To reach Spanish and Portuguese users in an efficient manner, ESA ground station of VILSPA (Villafranca del Castillo, Madrid) was prepared for the data reception and immediate transmission to DLR in Germany where pre-process is performed. Once at the EDG, located at LATUV (Valladolid, Spain), the algorithms are run and the products generated sequentially. The products were displayed automatically at user premises. All the processes were managed and monitored from INSA offices (Madrid, Spain).

112

Cameras Characteristics

WAOSS-B

MWiR

TIR

3.4 - 4.2 (MIR)

8.5 - 9.3 (TIR)

Wavelength

(\im)

0.60 0.67 (VIS) 0.84 0.90 (NIR)

Focal length

(mm)

21.65

43.39

43.39

Detector

CCD

CdHgTe

CdHgll

Ground pixel size (m)

185

370

370

Ground sampling distance (m)

185

185

185

Swath width (Km)

533

190

190

Table 1: BIRD sensors summary diagram On the other hand, LATUV facilities allows the direct reception of NOAA, FENG-YUN, T E R R A and M E T E O S A T satellites. Other high resolution data (LANDSAT) was bought at the end of the campaign for burned area assessment. Being this a demonstrator, internet is used for all the data transactions without incurring in long delays. For future experiments, several dedicated communication lines could be used instead.

i-

Processing for Products

BIRD images as obtained from D L R servers contain the calibrated NIR, MIR and TIR bands. The geo-referencing of the BIRD data was conducted at DLR using the BIRD on-board navigation data. In order to minimise the pre-processing time at DLR, it w a s decided to use the non coregistered and non geo-referenced BIRD Level 1 data for the fire recognition. Several products were generated attending to the phase of the forest-fire emergency, but along this paper, only the following types of products will be explained:

113

-

Products for fire detection tasks, delivering a near real-time hot spot map with parameter of fire outbreaks and already developed fires

-

Products for fire monitoring tasks, making operators able to see fire shape, location and released power, to perform a better management of the available resources

a) Fire Detection

Products

Two classical methods were used to recognise hot spots for the fire inventory map: threshold and contextual algorithms. Fixed-Threshold Detection Algorithms: the detection test analyses the brightness temperature values in the MIR and TIR bands and the difference of both (MIR-TIR). This scheme is found in the literature applied to different sensors like the NOAA-AVHRR, and it was adapted to BIRD payload Contextual Detection Algorithms: the surrounding pixel area (window) is used to establish thresholds according to the statistical characteristics (usually mean and standard deviation) of the background found in the window b) Fire Monitoring

Products

The monitoring products are maps of fire lines and associated parameters, but to obtain them is necessary to perform the detection process previously. Once the hot spot recognition is done, the next process is to determine the fire parameters: fire temperature, area affected and released energy intensity. Two levels of analysis have been distinguished to estimate these parameters, the analysis at sub-pixel level, where all the pixels affected are studied and characterised, and the analysis at cluster level. In order to carry out the analysis at a sub-pixel level, Dozier's technique was used, i.e. to solve a two-equation system set out for the radiance obtained by the sensor in the MIR and TIR bands: L, =τ, ρl(r,)+

(1 - p)LhiU,gl

+ρ L

m

i=

MIR,TIR

where L is the radiance of the pixel. Out of Dozier's equations, the fire temperature and the affected surface fraction will be obtained for each pixel and can be used to calculate the radiative energy release for hot pixels. A detailed description of the BIRD hot spot detection and analysis algorithm is given in (Zhukov et al, 2003).

Figure 5: Plume and fire flaming front using

114

BIRD

12«»o Figure 6: Hot spot map of 4" August 2003

)4«wo

Figure 7: Fire line maps of 4th August

1 1 ΟΟ»; JVAOEYJOO») ^OMOi S 0 0 0 3 OMSKS

/

fSOn

^ U F

,/FEE

h

Figure 8: Hot spots over Portugal l4

August

y

/

Ml··

"m t

D H

4

Figure 9: Fire line maps of 14" August

£

Campaign

Report

Fire inventory campaign was executed between the months of July and September 2003, with an initial test campaign performed the 30th of May. D L R and LATUV were the processing centres of BIRD data. The typical timeline of operations is as follows: Determination of the BIRD satellite site over-pass date and time by D L R BIRD team. G/S commanding by DLR German Space Operation Center (GSOC) Oberpfaffenhofen BIRD data acquisition and real-time down-link to the G/S (VILSPA and DLR-Weilheim, as back up) Data transfer from V I L S P A and/or DLR Weilheim to DLR Berlin Data pre-processing at DLR Berlin BIRD Level 1 data transfer to LATUV Geo-reference of data and RGB quick-look generation and transmission Fire inventory algorithm run and product served and archive updated Product management at user premises

115

09"' and 21st; August 2ld,

TÖh, 2fh and 2$h and September 1Ö'1

Results of the Hot Spots detected and the time from the satellite pass to the final information delivery are collected in Table 2 for all the campaign days. All the campaigns, except the algorithm validation test of the 30th of May, were run under real fire conditions. 4-

User

validation

Two methods were proposed for the validation process: in-field validation and off-line validation. The first was difficult and resource-consuming, taking into account the size of the region. The second, based on the comparison between the official fire reporting and the detected hot spots is more convenient but less accurate because some small fires may not be recorded and the position of many of them is only approximately calculated. To have a better understanding of the performance of the system, a test day was agreed with the fire managing end-user, who was kind to prepare a set of 15 prescribed burnings with different sizes in Galicia at the time of the satellite pass. The location of the fires was unknown for the project team. The campaign was then executed. Nine hot spots were detected, as summarised in Table 3 . A hot spot not initially in the list was reported, and minutes later, local authorities of Santiago the Compostela (La Coruna) confirmed the presence of a fire at that very precise location. For the rest of campaigns, there were occasions where a direct validation was possible but the unavailability of resources (fire fighting brigades were full time dedicated to extinction activities in larger fires) sometimes made it difficult to repeat this kind of validation exercise. For the bulk of the campaign an off-line validation was performed, but taking into account several considerations: Operators store fires using the central coordinates of the parish (village) containing the fire. Fires may be introduced in the database until 24 hours after their occurrence, only. A fire originated in one parish can be registered in another parish if boundary crossed. The geo-reference error of the algorithm is estimated between one and three kilometres. In general, 50% of the hot spots detected have a corresponding record on the database, but with a large error in the position, due to the sum of the above reasons. The user recognises the limitation of the manual fire reporting system and considers satellite fire inventory as an important improvement in the knowledge of the fire appearance and extinction cycles, sometimes

116

very difficult to follow w h e n local authorities and even volunteers take the responsibilities of first attack and suppression. Campaign Date

Tote! Delivery Time (mtn)

Hat Spots Detected

30415-2003

32

g

307

Hot Spots detected

2

23-09-2003

Orertse

BIRD d a b unavailable

Table 2: Results of the Fire Inventory campaign

4791814

14

59BOOO 4705000

Cloudy

CI

S74B96 4786025

Cloudy

Α

580061 4760877

C3

476190

579935

4758757

543926

4748004

25

20

2124

475G500

70

27-0S-2003

Cloudy S3B854

Fire Width (m)

32

25

50

10

to

1 5

01

613150

4683113

Cloudy

53

1.0 25

02

597713

4650524

Cloudy

48

OJ

550142

46574B9

Cloudy

25

1

04

563164

4685347

562010

4685751

1223

15

10.5

PI

7129

S49940

4666679

549872

4668807

100

9

Ρ1

583970

4742977

561137

4743166

517

59

0.75

Vi

544825

4682977

546324

46832?$

1529

25

Μ

551906

4592124

551246

4692835

1517

10

2

Table 3: Results of validation campaign (3(fh May 2003)

With respect to the fire monitoring products, generated off-line for Portugal, the validation has been performed using already validated satellite sensors, mainly MODIS. The cluster analysis shows the excellent radiometric performance of BIRD, whereas the definition of the fire lines is much better due to its better resolution in the thermal bands (from 3 to 5 times better). This is particularly useful to feed fire line propagation models, in general very sensitive to the input conditions. Portuguese users analysed the results admitting the good degree of matching with the available maps and the advantage of the satellite point of view w h e n large areas are affected.

4

Lessons

learnt

Satellite Earth Observation is a technology that opens a new range of possibilities in forest fire fighting. The challenge is to build full operational systems able to be embedded in current infrastructures and procedures. After the D E M O B I R D experience, operators are satisfied with the results and able to initiate pre-operational projects that make use of existing remote sensing data to better manage the fire fighting activities. S o m e conclusions can be extracted from the technical analysis of the mission, confirmed by the results of the s u m m e r campaign: -

BIRD channels are optimal for fire inventory and fire monitoring

-

Spatial resolution is appropriate for all the products

-

Radiometry of sensors is excellent

-

Data geo-referencing need to be improved

-

Re-visit time is precarious, and only acceptable for a demo exercise

-

Duty cycle is a weak point; however, during this demonstration, availability was very good

-

Data delivery is quick and reliability acceptable

Summarising, an excellent sensor was built on a quite limited platform, which should improve the pointing knowledge capability and the available power; the ground segment is acceptable although dedicated lines could improve delivery time. F U E G O S A T dedicated mission should solve the aforementioned shortcomings to provide an excellent product that fully satisfy user community. Fire inventory product was labelled as excellent by users. The system was quick and very sensitive. Fire monitoring product offered very promising and impressive results w h e n tested over large fires. This product allows fire-fighting managers to better understand fire characteristics, shape, position, status and future evolution. Unfortunately, during this demo it was not possible to run a real-time exercise over the large events of Portugal.

117

Validation of the products became very difficult due to the work load of fire-fighters during the season and the uncertainty of the fire position (given by satellite platform limitations). The validation in a real scenario is always difficult but a better geolocation would help to reach the fire area more quickly. 500 m error is recommended. Large fires are not always cartographied in real time but once fire is extinguished. Fire monitoring product can only be validate at the end or by using indirect methods like other satellites. This occurred in Portugal (huge fires) last summer. Finally, user databases were available to the team in a 'read-only' fashion. This was reasonable agreed in order not to jeopardise critical data included in those datasets. For future projects, the satellite can be a source of new database records, the same as currently the operator or an external phone call can be. User is able to accept this since the system is known and the space capabilities have been demonstrated. Of course, this will require an extra implementation effort.

J-

Conclusions

After performing DEMOBIRD demonstration during 2003 summer campaign, BIRD satellite sensors have probed to be excellent for fire inventory and monitoring, and good enough for burnt area generation, as operators have confirmed. It is possible to generate a whole catalogue of products based on BIRD data, but the main problem is the re-visit time, only acceptable for demonstration situations. BIRD have some minor deficiencies, such as a bad georeferencing, problem that can be solved easily, but increasing final product processing time. To increase re-visit time, it is suggested to make wider the spatial segment, to include images from other satellites. The idea is to supply Fire data monitoring products twice a day for a near full operational cycle. Another important aspect to perform this near operational demonstration is that delivery times need to be minimised. This can be achieved by centralizing process in a specific centre, where data reception and data processing can be done. Another interesting suggestion is the idea of performing a mobile demonstration, that is, to integrate a set of space technologies in a vehicle to perform data download and transmission to operators of the regions where fires have been located. Satisfactory DEMOBIRD results have opened the way to develop near full-operational demonstrations based on space technology. The experience has supplied a set of ideas from operators that will have to be considered in the future to attract operators attention over these kind of technologies in order to develop a full operational service.

ι

References

Briess, K., Jahn, Η., Lorenz, Ε., Oertel, D., Skrbek, W., Zhukov, B., 2003, Fire recognition potential of the bi-spectral Infrared Detection (BIRD) satellite, Int. J. Remote Sensing, Vol. 24, No 4, 865-872 Gonzalo, J., 1998, A low cost service for fire detection. Proc.3rd Int. Conf. on Forest Fire Research, Portugal.

page 2029. Luso,

Lorenz, Ε. and Skrbek, W. 2001. Calibration of a bi-spectral infrared push-broom imager. Proceedings of SPIE, Infrared Spaceborne Remote Sensing IX, San Diego, 29 july-3 August 2001, 4486, 90-103 Skrbek, W., & Lorenz, Ε. 1998. HSRS - An infrared sensor for hot spot detection. Proceedings borne Remote Sensing VI Vol. 3437, 167 - 176.

of SPIE, Infrared

Space-

Wooster, M. J., Zhukov, Β & Oertel, D. 2003. Fire radiative energy for quantitative study of biomass burning: Derivation from the BIRD experimental satellite and comparison to MODIS fire products. Remote Sensing of Environment. Vol. 86, 83-107. Zhukov, B., Briess, K., Lorenz, Ε., Oertel, D. & Skrbek, W. 2003. BIRD Detection and Analysis of High-temperature Events: First Results. Proc. SPIE Remote sensing for environmental monitoring, GIS Applications and Geology II, Vol. 4 8 8 6 . 1 6 0 - 171

118

RESULTS OF SPACE WEATHER RESEARCH ON SCIENTIFIC-EDUCATIONAL MICRO-SATELLITE "KOLIBRI-2000" S. I. Klimov 1 ' 2 , O. R. Grigoryan 2 ' 3 , V. A. Grushin 1 2 , D. L.Novikov 1 , V. L.Petrov 3 , S. P.Savin 1 1) Space Research Institute of RAS, Profsoyuznaya 84/32, 117997 GSP-7, Moscow, Russia, Phone: +7 095 333 1100, Fax: +7 095 333 1248, [email protected] 2) Interregional public organization "Microsputnik", Profsoyuznaya 84/32, 117997 Moscow, Russia; Phone: +7 095 333 3077, Fax: +7 095 333 3077, [email protected] 3) Institute of Nuclear Physics, Moscow Sate University, Vorob'evy Gory, 119992, Moscow, Russia, Phone: +7 095 939 5061, Fax: +7 095 939 0896, orgri@,srd.sinp.msu.ru ABSTRACT The results of space studies are used in many applications, including the education. Work with the schools is a natural method to inform the general public about the role and value of space studies for humanity. The first Russian- Australian scientificeducational micro-satellite "Kolibri -2000" of total mass of 20,5 kg, on 20 March. 2002, has been injected into orbit of International Space Station (ISS) by separation from the transport vehicle "Progress". It began the development of tasks for scientificeducational micro-satellite (SEM). In spite of small size, SEM had 3.6 kg of scientific payload, which provides an opportunity to carry out rather wide scientific studies both in the field of "classical" space physics and for the space weather, atmosphereionosphere connections etc., it serves also for the tasks of space education. According to the preliminary ballistic calculations, "Kolibri-2000" had to fly about 4 months; however, 17-20 April of 2002 the rapid reduction of its height began, first of all due to the increase in the Sun activity. In this paper we address the influence on the ionosphere of the processes, which occurred on the Sun of 14-24 April of 2002; changes in the flows of energetic particles, magnetic and electric fields has been examined 1. INTRODUCTION The "Kolibri-2000" micro-satellite (MS) [1, 2] has been designed as a base satellite for the entire Program of Scientific - Educational Micro-Satellites (PSEMS' 2002-2007) [3]. The program of scientific researches was realized by the usage of the complex scientific equipment (CSE) aboard "Kolibri-2000". including: The fluxgate magnetometer (TFM) [4] to measure intensity of Earth magnetic field three orthogonal components (Bdcx. Bdcy, and Bdc) in the range +/- 64000 nT and of its fluctuations in frequency band 0.01-15 Hz (Bacx. Bacy, and Bacz) in the range +/100 nT, along with spectral density of fluctuations in the band 50-60 Hz (B50) for one component. TFM data were used also by onboard magnetic-gravitation system of orientation (BUSOS), served for damping of MS fluctuations [5]. The analyzer of particles and fields (AChP) [6] to measure: a) Precipitated particles and rays: electrons with energy > 75 keV (in two directions - zenith and nadir); electrons with energy > 300 and > 600 keV (in

119

different direction along an axis, perpendicular to zenith}; protons with energy > 50 MeV; neutrons of 0.1 eV - 10,0 MeV; γ-rays with energy: > 50 keV and > 2 MeV. b) Electric field: DC field (Edc) one component in the range +/- 2560 mV/m; spectral density of fluctuations in the frequency band of 50-60 Hz (E50) for one component. It has been shown [7] that the difference between the flow of energetic electrons in the quiet geomagnetic conditions (on 7 April) and in the disturbed ones (on 22 April) can reach 5-6 times. Measurements on board the satellites with small mass make it possible to increase noticeably the accuracy of monitoring the parameters of cosmic radiation, especially neutron emissions. This is because of during the measurements on board the massive space objects, such as ISS (>100 tons), the significant contribution in the of detector's count rates result from the secondary radiation, which is produced by interaction with the spacecraft of the energetic charged particles of the galactic cosmic rays and that of the Earth radiation belts. 2. ELECTROMAGNETIC CLEANLINESS. Data processing during the flight and their physical interpretation have shown, that the approach to electromagnetic cleanliness, realized during design and tests of the "Kolibri-2000", provided a high-level of electromagnetic cleanliness in flight. It has ensured the measurement of magnetic and electric fields with sensitivity, exceeded in the appropriate ranges the levels, achieved in a number of the specialized international projects, for example Interball-1 (Fig. 1) [8].

1 Frequency (Hz)

a) Electric field. Curve 1 , 2 - Interball-1 and Prognoz-8 bow shock crossing; Curve 3, 6 — Interball-1 background spectra in solar wind; Curve 4, 5 - Interball-1 filter-bank spectra at bow shock.

10

i 00

1000

b) Magnetic field. Curve 1 - bow shock spectrum from the Prognoz-10; Curve 2 Interball-1 bow shock spectrum (FFT of waveform); Curves 3,5 - Interball-1 noise spectrum (FFT of waveform and filter-bank spectra); Curve 4 - Interball-1 bow shock.

Figure 1. Comparison of sensitivity of electric and magnetic field measurements on "Kolibri-2000" (crossed circle), Prognoz-8, Prognoz-10 and Interball-1.

120

100C

Frequency (Hz)

It should be noted that the sensors of electric field (cylindrical films with diameter 20 mm and length 100 mm) were mounted directly at the ends of the solar panels, and the fluxgate sensor for magnetic field has been placed inside the MS body. 3. MAGNETIC STORM STUDY. Dst index is generally used to scale the strength of a magnetic storm. It results from perturbations of the Earth's magnetic field generated by the ring current. Two doublepeaked intense magnetic storms occurred 20 and 23 April 2002. 3.1. Wave Measurements Because of the low-noise level aboard MS it was possible, in particular, to study of some individual events from three component measurements of magnetic field in the range of 0.1-15 Hz. Such measurements on "Kolibri-2000" have been carried out on 22 April (19:54 - 20:20 UT) and on 24 April (13:30-14:01 and 14:20-15:25 UT). Data on 22 April are presented in Figures 2 and 3. „ k. X_ I i -i --ii-J I — •L :a_ Λ ι I ä Γ· : J* _2a_ f $ ψ a_i s13 τϊ _ζ Τ L '•--? Ψ. 3 Έ -> Γ i Μ F • »J 3 ΐί' •JL iE ; Ζ L t L 2 1 ui '.Γ Β 'sf ^ i Λ α«ϋ K!95 - ί I ffrrt kΝζ -S 1 s ε*. —. 1 ί % τΠ V b & J Λ : . i M i Iii lit ä k ί ir fTf1 11; .1A'· • t i" /r / Λ — " Ü . 7" 1 J_ J KL / ia a_i J r s h iii ύΊ it Ii Ί 'J IΛ; Γ\ _LLL Jt JI • C \ Γ < 1 . tfl * Q • u ζ j 1 h. toiüaaiL ILk HJύ LL:i iL·.. L· L ZflΛ fei Ια ifl £'±ns j ΪΟ. L i i • Zfl k] ja ICJjjv ; LiX!b ÜJ ih _LL L \ LLL.L 2 „ 1 L O _ TO π Γϋ V •J ^Ll J_ L Ii ... _ \ -J Figure 2. Wave form Bacx, Bacy, Bacz and orbit 22.04.2002 (19:55-20:18 UF).

,,

\

%

|

2

54

19:57

20: 0

20: 4 20: 7 T I M F IJT

20:10

2 0 : 14

Figure 3. Magnetic and electric field measurements at 22 April 2002. From top. Bacx, Bacy, Bacz - A C magnetic field in the range 0.01-15 Hz [IV =40 nF] (dotted lines) & Bdcx, Bdcy, Bdcz - DC magnetic field

20:1 7

-·.

2

1 2.8

22.1

31.7 41.3 50.9 INVAR Μ I AT

59.2

03.2

jL.«? 9.4 1 5

,

10.5

ΜIΤ

,

12.3

,

14.9

1 .0

0.5 0.0

- 0 .5 3 1 9.0 322 7 Al T I T U r i F .

[1V = 25600 nF] (dashed dotted lines); B50 - spectral density of magnetic field fluctuation in the range 50-60 Hz [ I V = 3 nF]; E dc2 - DC electric field [ I V = 160mV/m],

O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O

Ο - sun X - shadow

Figure 2 presents the parallel data about the wave form on three components in the united scale and in the projection on the position of orbit MS in the Mercator

121

coordinates. Difference in the intensity of fluctuations on components AC magnetic field is predominantly connected with different orientation of the components of field relative to the vector Β of the intensity of Earth magnetic field and also with fluctuations MS relative to vector, which is reflected in Figure 3. Interesting is the fact of the observation of intensive fluctuations on the component Bacy with the approach to the North American coast in the morning local time. In this case the measuring component Bacy of is directed along the vector B. Unfortunately, this is the only case of measurements "Kolibri-2000" in such type of orbit.

b) Bacv

c) Bacz

Figure 4. Wavelet square modulus of Bacx, Bacy, Bacz 22.04.2002 (19:55-20:18 UT). (intensive signals at the lowest frequencies in the beginning and the end are effects of wavelet analysis)

122

The more detailed analysis of the nature of the fluctuations, represented in Figure 2, is investigated with the aid of wavelet the spectral analysis, whose data are presented in figures 4a - 4c. First of all, it is necessary to note that in the spectral diagrams are not observed the emissions at the fixed frequencies, which, as a rule, are the electromagnetic interferences. This also confirms the high electromagnetic cleanliness of '"Kolibri2000". In the data, represented in figures 4a - 4c clearly are separated several events. In the beginning the event, conditionally noted by interval of 1—1, is separated. Event 1—1 (19:57-20:02 UT) is characterized by the decrease in the "central" frequency from ~ 1 Hz to 0.5 Hz on the component Bacx and from 0.5 Hz to 0.01 Hz on the component Bacz. The decrease in the frequency occurs approximately with the same inclination (dispersion) that, as a rule, it testifies about united nature of signals. It is important to note that the decrease in the frequency is not observed on component By and fluctuations in these frequency bands practically be absent. Taking into account the fact that components Bx and Bz in this event predominantly are located at large angle to vector Β (the DC magnetic field is close to the plane X-Z). it is possible to make a conclusion about that the transverse wave packet has nearly circular polarization with frequency of geomagnetic pulsation Pel. Measurements of both magnetic and electric components (Figure 3) of geomagnetic pulsation aboard a spacecrafts have been done sufficiently rare [9, 10]. "Kolibri-2000" data on 22 and 24 April are unique for the lowest ionosphere at the height of ~ 320 km. In the frame of the interval 1—1 it is observed short event (-20:00 UT) with more complicated spectra like "inverted V". Spectrum in component Bacx during 20s displays interchanged maxima at ~ 0.2 Hz and 1. Spectrum in component Bacz at this time has maximum, gradually displacing from 0.8 Hz to 0.15 Hz. During event 1—1 and in spectrum of DC (< 0.5 Hz) electric field the intensity of fluctuations at frequencies 0.02 - 0.08 Hz by two orders of magnitude exceeds that of 0.1 - 0.5 Hz. For more precise interpretation of this event it is necessary, to restore the orientation of the spacecraft axes, which are fluctuating relative to Earth's magnetic field. Event 2—2 is observed only on the component Bacy is apparently the same process, as event 1—1, fixed when the measuring Y axis is located at large angle to the vector B. Event 3—3 - wide-band (from 0.5 Hz to 5.0 Hz) emission in the nature is close to event 1—1, but without the dispersion. It is important to note that events 1—1 and 2—2 were observed close to the geomagnetic equator (fig. 3; INVAR M.LAT < 15 and ML Τ ~ 8-9), while the event 3 - 3 - on middle latitudes (INVAR M.LAT ~ 45 and MLT ~ 09-12). Event 4—4, first of all, is characterized by the fact that the component X long time is perpendicular to vector Β (Fig. 3) and components Y and Ζ are turned to 180° in the plane, which contains vector B. Here, on the approach to the sub-auroral latitudes, only on Bacy (Fig. 4b) are observed several wave packets with a changing carrier frequency of from 0.1 Hz to 0.8 Hz and then the decrease to 0.05 Hz. It is frequency of geomagnetic pulsation Pel. Event 5—5 practically repeats event 3—3 from a somewhat increased intensity. In interval of 5 is observed event 6—6 in the form of wave packet with a carrier frequency of 0.2 Hz. The large intensity of fluctuations is certainly connected with presence MS in the sub-auroral zone.

123

3.2. Wave and Energetic Particles Measurements For the more precise physical interpretation of wave data was carry out the analysis together with the data on the energetic particles, measured by instrument AchP [6] on 24 April 2002 (13:30-14:02 UT). The geographical location of orbit the "Kolibri-2000" is represented in Figure 5, where is also given the wave form Bacy. In the geophysical coordinates the orbit is represented in Figure 6. Kolibri; R *

?4-n4-?ün?

π fi rlo-ri I m a p n

•gl 1 Ό. A. S %

--

—

V

Ν r

Ί

1 A-:

^

ί


,) is a sign of derivative of the component of the geomagnetic field alongwith the axis of symmetry of the satellite. For practical algorithm it is necessary to determine a sign of difference between measurements of the magnetometer alongwith the axis of symmetry at two consecutive moments of time. Lhe idea of the suggested control algorithm for the coil lies in imitation of action of the nutational damper designed with usage of a soft-magnetic material. Remagnetization of the material causes a dissipation of energy of the nutational motion. Lhe control algorithm simulates the magnetic moment of idealized hysteresis rod with a rectangular hysteresis loop. 5. NUMERICAL SIMULALION OF ALLILUDE MOLION In this section we present the results of the numerical simulation of the attitude motion of the satellite with orientation system uses two different algorithms. Lhe first simulation was carried out for the first algorithm with alternate use of control laws for the satellite orientation described by formulae (2), (5) and control law for the nutation motion damping described by the formula (9). Alternate use of two control laws was carried out in following way. Firstly, during the certain time interval the control law for the satellite orientation was working then during other time interval control law for nutation motion decreasing was activated. We pay attention that the combination of the control laws requires information about current angular velocity of the satellite. For numerical simulation we use the following parameters of the satellite orbit: longitude of the ascending node Ω, = 0 , inclination of the orbital plane ί = π!2, parameter ρ = 7.02e+6 m, eccentricity e - 0 . Maximum dipole moment of the coils is '"max = 1 A-m2, the tensor of inertia is J = diag(0.09,0.09,0.18) kg-m2, a required spin axis direction coincides with a normal to the orbital plane. Fig.3 shows time dependence of the angle between the required direction and current direction of the symmetry axis. Lhe time dependence of three components of angular velocity is shown in Fig.4. Lhe results are obtained for the following initial conditions: the axis of symmetry lies in the orbital plane, initial angular velocity © = (0.5,0.5,0) sec"1. We consider the modeling case when orientation and angular velocity of the satellite are known at any time. Lhe next simulation was carried out for the second algorithm. For numerical simulation we use the orbit parameters, the parameters of the satellite and the initial

342

conditions given above, except the capacity of magnetic coils. In this case we modeled four coils: two coils (big and small) located along symmetry axis of the satellite with capacities 0.8 A-m2 and 0.2 A-m2, and two perpendicular coils located in an equatorial plane of the satellite with capacity 0.1 A-m . The satellite orientation is carried out in three stages. At the first stage with duration 7j = 5000 sec. the components of angular velocities ωγ and ω2 were decreased using the control law (14). At the second stage with duration T-, = 8000 sec. the satellite was rotated under the control law (13). The small coil was used for the nutational motion damping. The big coil was switched off. The third stage proceeds during all time of satellite functioning. At this stage the satellite orientation was made out using big coil under the control law (12). The small coil was also used for the nutational motion damping. Fig.5 shows time dependence of the angle between the required direction and current direction of the symmetry axis. The time dependence of three components of angular velocity is shown in Fig.6.

p

i

^

2000

10000 12000

400Q ,

600Q t (sec)

2000 _,

4000 ,

6000 "SSM

10000

12000

2000

4000

6000 1 Issel

10000

12000

^Im*·" 2000

4000

6000

10000 12000

t (sec)

Fig.3: Time dependence of angle a between the required direction and current direction of the axis symmetry (first algorithm)

Fig.4: Time dependence of three components of the angular velocity (first algorithm)

Fig.5. Time dependence of angle a between the required direction and current direction of the axis symmetry (second algorithm)

Fig.6. Time dependence of three components of the angular velocity (second algorithm)

343

6. SUMMARY Passive and active MACS described above are realized for the first two 7W5-series nanosatellites which are developped by the Russian Research Institute of Space Device Engineering (RNII KP) in collaboration with the Keldysh Institute of Applied Mathematics of RAS. The TNS-0 with passive MACS (Fig.7) is launched from ISS on of March, 2005. Mock-up of TNS-1 with active MACS is shown in Fig.8.

tMg.8. Mock-up ot TNS-1

(courtesy RNII KP) 7. AKNOWLEDGEMENTS The work is supported by the Russian Basic Research Foundation (Project Ν 03-0100652), the Russian Research Institute of Space Device Engineering and by the Federal Agency of Science and Innovations of Russia (Project Ν 02.700.12.050), NATO Fellowship (2004). 8. REFERENCES [1] [2] [3] [4] [5] [6] [7]

[8]

Renard M.L., Command Laws for Magnetic Attitude Control of Spin-Stabilized Earth Satellites. Journal of Spacecraft and Rockets, Feb. 1967. v.4, N2, pp. 156-163. Wheeler P.C. Spinning Spacecraft Attitude Control via the Environmental Magnetic Field. Journal of Spacecraft and Rockets, Dec. 1967, v.4, N12, pp 1631-1637. Ergin E.I., Wheeler P.C.. Magnetic Attitude Control of a Spinning Satellite. Journal of Spacecraft and Rockets, 1965, v.2, Ν 6, pp.846-850. Sorensen J.A., A Magnetic Three-Degree of Freedom Attitude Control System for an Axisymmetric Spinning Spacecraft, Journal of Spacecraft and Rockets, 1971 v.8, N5, pp 441-448. Shigehara M, Geomagnetic Attitude Control of an Axisymmetric Spinning Satellite, Journal of Spacecraft and Rocket, June 1972, v.9, N6, pp 391-398. Ilyin A.A., Ovchinnikov M.Yu., Penkov V.l., Orientation Maintenance of the Small Spin-Stabilized Satellite, Preprint of HAM RAS, 2004, Ν 83, 28p. Ilyin A.A., Ovchinnikov M.Yu., Penkov V.l.. Selivanov A.S., Magnetic Attitude Control System for the Russian Nano-Satellite TNS-1, Paper IAC-04-A.3.10, 55,h Congress IAF, 4-8 Oct., 2004, Vancouver, Canada, 10p. Barabash S., Kiryushkin I., Norberg O., Ovchinnikov M., Penkov V., The Nanosatellite Munin, a Simple Tool for Auroral Research, Advances in Space Research, 2003, V.31, Issue 2, pp.313-318.

[9] Ovchinnikov M., Pen'kov V., Passive Magnetic Attitude Control System for Nanosatellite, Cosmic Research, 2002, V.40, Ν 2, pp.156-170.

344

ANALYSIS OF THE HAUSAT-2 ATTITUDE CONTROL WITH A PITCH BIAS MOMENTUM SYSTEM Young-Keun Chang, Byung-Hoon Lee, Soo-Jung Kim Space System Research Lab., Hankuk Aviation University Goyang, Gyeonggi-Do, Korea Phone: +82 2 300 0286, Fax: +82 2 3158 3189, [email protected] ABSTRACT The HAUSAT-2 is a 25kg class nanosatellite which is being developed by graduate students at the Space System Research Laboratory (SSRL). The HAUSAT-2 implements a pitch bias momentum method for 3-axis stabilization, consisting of one momentum wheel in pitch axis and 3-axis magnetic torquers. 2-axis sun sensors, 3-axis magnetometer, and star tracker are used to determine the attitude. This paper describes two primary attitude control modes and control performances to perform the following functions: 1) initial acquisition (initial attitude acquisition mode), 2) 3-aixs stabilization (mission mode). For detumbling the satellite, the initial mode uses a momentum wheel and magnetic torquers with B-dot control law. The mission mode also uses a momentum wheel and magnetic torquer to maintain 3-axis stabilization. The control laws used during these control modes are analyzed. Especially, magnetic torquer's control law is implemented with Pulse-Width Modulated(PWM) onoff control, and shown to work equally well with the original continuous and variable strength control law. Also, an attitude control method using only 3-axis magnetic torquers was also considered for HAUSAT-2 increase system reliability, in preparation for a momentum wheel failure.

1. INTRODUCTION The HAUSAT-2, currently under development by Space System Research Lab., is a nanosatellite that will fly a sun-synchronous orbit at low earth orbit (LEO). Its primary mission objectives are to provide studying the scope of activity and ecology of animals using Animal Tracking System (ATS) and collecting the space information data of mission orbit from Electric Plasma Probe (EPP) as a space science payload. Secondary mission objective is to test the performance of the star tracker which is the first development in KOREA. In order to accomplish missions, an earth-pointing performance of less than ± 1° in pitch and ± 3° in roll and yaw axis is required. Pitch bias momentum method has been chosen in consideration for the operation of a star tracker, which requires higher pointing accuracy. In the Attitude Determination and Control System (ADCS) configuration, The HAUSAT-2 is stabilized to earth pointing by a momentum wheel(MW) aligned with the pitch axis and 3-axis magnetic torquers (MTQs). A MW is nominally spinning at a particular

345

rate and change speed, i.e., angular momentum provides stability to the wheel axis which is perpendicular to the orbit plane and the torque generated by wheel about the wheel axis is used to stabilize the attitude of the satellite in the orbit plane. But, the momentum bias is not itself sufficient to adequately stabilize the attitude in the roll and yaw axes. MTQs are used to assure an accurate attitude stabilization, keeping the rollyaw attitude errors within small range. [1-3] Attitude control consists of two operating modes: initial attitude acquisition mode and mission mode. The initial attitude acquisition mode is an attitude control mode for detumbling the satellite. The mission mode is the mission execution mode for HAUSAT-2 stabilization. In this mode, 3-axis stabilization is achieved by attitude control. In this paper, an appropriate operation method and performance verification for the actuator according to each satellite operation mode have been verified, when pitch momentum bias method is implemented for the HAUSAT-2 attitude control and stabilization.

2. HAUSAT-2 ADCS DESIGN CONCEPTS AND REQUIERMENTS The HAUSAT-2 is a 25kg class nanosatellite and is limited in size, weight, and power compared to other generic satellites. Accordingly, power consumption and mass become the biggest factors in designing a satellite attitude control system. The allocated budget for HAUSAT-2 ADCS is 4kg for mass, and 6W for average power. The sensors and actuators design must fit under this budget limit. ADCS is divided into sensors that determine the attitude, and actuators that correct the attitude. Since HAUSAT-2 is a small satellite that suffers from more stringent limitations on mass and power, simple yet reliable parts were used in the development. The HAUSAT-2 uses sun sensors and magnetometers for the sensors, and MW and MTQs for the actuators. Sun sensor is consisted of four fine sun sensors and one coarse sun sensor. Fine sun sensors were mounted at each side panel for attitude determination and coarse sun sensor is positioned at anti-nadir panel for determining the direction of sun. Magnetometer includes magnetoresistive sensor because of mass and power limitation. Magnetic torquers are constructed in-house from coils of copper magnet wire covered with epoxy for stiffness, and then wrapped in kapton tape. Each magnetic torquer coil is capable of producing a magnetic moment of 2Am2. The momentum wheel will be placed in the pitch direction, to provide the satellite with inertial stiffness in the orbit normal direction. Figure 1 shows ADCS components on HAUSAT-2, and Table 1 lists the component specifications.

Figure 1. ADCS Components and Coordinates

346

Table 1. Specification of ADCS Components Mass

Power

Performance

Sun Sensor

lOOg(lEA)

O.IW(IEA)

±0.5°

Magnetometer

75 g

0.3W

±30nT

Star Tracker

2 kg

6 W

10 arcsec

Magnetic Torquer

270 g (1EA)

1W(1EA)

2 Am 2

Momentum Wheel

1kg

2.1W

0.09 Nms

3. HAUSAT-2 ATTITUDE CONTROL PERFORMANCE Spacecraft in orbit encounters disturbance torques from various environmental sources. These torques are accumulated over time at the spacecraft body. Different environmental torques are more prevalent at different altitudes.[5,6] The HAUSAT-2 will stay in a sun-synchronous orbit at 650km altitude. In low earth orbits, the largest environmental torques are gravity-gradient, magnetic, and aerodynamic. In this paper, gravitygradient, magnetic and aerodynamic torques are considered. These disturbance torques are shown in Table 2. Magnetic control of spacecraft is dependent upon the local magnetic field. For modeling the MTQs' control torques, mathematical models of the geomagnetic field using the International Geomagnetic Reference Field (IGRF) are used in this paper.[3,7] Table 2. Disturbance Torques Type o f disturbance

Max. torque (N-m)

Magnetic field

2.293 X10" 6

Gravity field

0 . 1 5 4 X 10"6

Solar pressure

0.0586X10" 6

Atmosphere

0.0688 X10" 6

3.1 Initial Attitude Acquisition Mode The initial acquisition mode is an attitude control mode for detumbling the satellite. This mode is also used when the satellite is lacking mission requirements, onboard com-

347

puter resets, or an anomaly has been detected during the mission phase. B-dot logic is used in this mode, as generally used for controlling initial angular velocity.[l,2,4] B-dot logic defines each axis dipole moment as follows. Mi=-KBi

(i = 1,2, 3),

K> 0

(1)

where, Κ is the control gain constant, B: is a differential of i-axis earth magnetic field value. Bj will be calculated directly by differentiating the magnetometer output data. The MTQ's control torque on a dipole in a magnetic field Β is given by equation (2). Tmtq=MXB

(2)

In fact, MTQs are activated in an on-off mode. So we used Pulse Width Modulation (PWM) method to generate dipole moment by on-off mode.[5] A PWM control timing is applied to the each axis as shown in Figure 2. In the case of MTQ_1 which produces roll axis dipole, for example, the output data of magnetometer is taken during 0.25 second in the beginning, and then MTQ_1 is operated by the PWM method to produce the necessary torque within the maximum 10 pulses during the last 0.75 second. At this time, the interval of a pulse is set to 0.25 second. Every 4 second, control torques are applied to each axis in the similar way.

Magnetometer

MTQ_1

Magnetometer

MTQ_1

Magnetometer

Figure 2. Example of PWM Control Timing Figure 3 shows the results in which the M W was spun up before separation. The initial orbit referenced angular rate is set to 2°/sec about all axes. Once the nutation is damped out, reorientation of the spin axis begins. As shown in Figure 3(b), angular velocity stabilizes to about twice the orbital rate and attitude is stabilized with the pitch axis making a 40° angle with the orbit plane in Figure 3(a). Therefore, when the angular velocity falls within a predetermined range for mission phase, the attitude control mode changes to a mode for mission execution (mission mode). Figure 4 shows the results of initial attitude acquisition mode using the MTQs only in the same initial conditions. In this case, angular velocity stabilizes to about twice the orbital rate, but attitude is not controlled.

348

,

a) Orbit normal/Pitch-axis Error angle

b) Angular Velocity

ΜΤ

0

(3)

and Kd are control gain constants, and Twv is the wheel control torque.

MTQs can only generate torque in the perpendicular direction to the magnetic field vector Β . In order to avoid interference with the pitch axis controlled by the MW, only the MTQ mounted on the pitch axis is used to control roll and yaw. In this case, the pitch axis dipole moment created is as follows.[1,4]

M2 = -ΚλΒλφ-K2B2

Kt,K2 > 0

(4)

where, Κλ and K2 are control gain constants, Bl is the earth magnetic field with respect to roll axis, and B2 is the rate of change in earth magnetic field with respect to pitch axis. Figure 5 illustrates the simulation results. As shown in Figure 5(a), HAUSAT-2 pointing precision requirement of less than 0.5° in pitch axis is satisfied. The MW is the most important actuator for the HAUSAT-2 that utilizes pitch bias momentum method. In general, attitude control reliability of MWs is guaranteed by having a backup wheel in case of an anomaly in the main MW. However, such design cannot be accommodated in nanosatellites such as HAUSAT-2 due to design limitations. This is why an attitude control method using only 3-axis MTQs was also considered for HAUSAT-2 for increase in system reliability, in case of a MW failure.[5,6] In this case, the MTQs produce dipole moment around each axis as described below, in order to stabilize the satellite.[3] (Bxu,)

Μ

(i = 1 , 2 , 3 )

(5)

Figure 6 shows the simulation results using MTQs only. As the results of Figure 6, it can be seen that it takes a long time for the stabilization due to low level of torque capacity. However, a pointing accuracy, which is required for the HAUSAT-2 mission accomplishment, is fully satisfied.

350

Ο

1000

2000

0000

Ό00

GOOO

COOO

7000

\

a) Orbit normal/Pitch-axis Error Angle

0 .,,-51000

2000

3000

4000

5000

6000

7000

8000

9000

0

2000

3000

4000

5000

6000

7000

8000

9000

1000

0000

3001

-

b) Attitude

c) MTQ Control Torque

d) MW Control Torque

Figure 5. Mission Mode (MW + MTQs)

h

£000

\

Λ

·

uu

.uuu

UU uu

uu

"V 2000

4000

6COO

8000

10000

. 12000

14000

16000

-

^

18000

a) Orbit normal/Pitch-axis Error Angle

b) Attitude M-ij„i I

/ v

^

,-32000

4000

6COO

8000

10000

12000

14000

16000 | —

18000

2000

4000

/

Λ

,-e2000

40DO

6000

800D

1000C

12000

K000

16030

18000

Μ

M= \

/ 0

18000

'»,„

-

\

0

160DO

I —

6COO

8000

10000

12000

14000

15000

18000 —

\ V c) Magnetic Dipole Moments

ίντο^ I

— - -

d) MTQ Control Torque

Figure 6. Mission Mode (MTQs only)

351

4. CONCLUSION The HAUSAT-2 implements momentum wheel and magnetic torquers as actuators for attitude control, as well as pitch bias momentum control method. It has been shown through simulations that this method satisfies the initial attitude acquisition and pointing precision mission requirements. In addition, mission execution using only the MTQs has also been analyzed in case o f a M W failure, and such performance also meets HAUSAT-2 attitude control requirements. It has been verified here that pitch bias momentum method is a control method applicable to HAUSAT-2. A more detailed verification analysis that takes into account attitude control method for momentum dumping and sensor noises is being performed.

5. ACKNOWLEDGEMENT This research has been supported by "National Research Lab." Contract awarded Korean Ministry o f Science and Technology. We would like to thank for the financial support. 6. R E F E R E N C E S [1] Stickler. A. C. and Alfriend, Κ. T „ "Elementary Magnetic Attitude Control System," Journal o f Spacecraft, Vol. 13, No. 5, 1975, pp. 282-287 [2] Alfriend. Κ. T„ "A Magnetic Attitude Control System for Dual Spin Satellites," AIAA Journal. Vol. 13, 1975, pp. 817-822 [3] Makovec. K. L.. "A Nonlinear Magnetic Controller for Three-Axis Stability o f Nano Satellites," M S Thesis, Virginia Tech., 2002 [4] Whitford, C. and Forrest, D., "The C A T S A T Attitude Control System," 12th AIAA/USU Conference on Small Satellites. 1998 [5] Sidi. M., "Spacecraft Dynamics and Control," Cambridge University Press, 1997 [6] Chetty, P. R. K „ "Satellite technology and its applications," McGraw Hill, 1991 [7] Wertz, J. R., "Spacecraft Attitude Determination and Control," Kluwer Academic Publisher, Dordrecht, 2000

352

A MICRO SUN SENSOR FOR EARTH-OBSERVATION NANOSATELLITES FLYING IN FORMATION Giancarlo Rufino, Michele Grassi, Vincenzo Pulcino DISIS - Dept. of Space Science and Engineering "Luigi G. Napolitano" University of Naples "Federico II", P.le Tecchio 80, 80125 Naples, Italy Phone:+390817682159, Fax:+390817682160, [email protected] Alexander Degtyarev Staff Scientist, Keldysh Institute of Applied Mathematics RAS, 4 Miusskaya Sq., Moscow, 125047, Russia

ABSTRACT This paper deals with the validation of a micro sun sensor under development at the University of Naples. Sun line determination is performed by the centroiding of sun images on the focal plane obtained with a multiple-hole tiny mask. Sensor laboratory calibration is performed with neural-network-based techniques using a home-made test facility. Processing of 1-hole data demonstrates sensor precision better than 1-arcmin. Preliminary results with multiple-hole data show that sensor precision can be improved by averaging multiple sun-images.

1. INTRODUCTION Over the last few years distributed architectures have received increasing attention. Technological feasibility of distributed architectures relies also on the increasing reduction in system mass, size, and power requirements [1]. Indeed, satellite close and coordinate flight poses demanding requirements on satellite relative navigation and control. Concerning navigation, operation autonomy can be achieved by the fusion of several sensors' outputs on board the satellite. In this view, the use of MEMS-based (Micro-Electro-Mechanical-Systems) sensors is a promising solution to reduce mass and power requirements. Concerning electro-optical sensors, Active Pixel Sensor (APS) technology is playing a dominant role, since it allows on-chip A/D conversion and integration of detector and electronics into a single chip. Latest generation star trackers and sun sensors adopting such components are being studied and realized [2,3], Recently, the Italian Space Agency (AS1) has started a short-term program for the development of a high-technology microsatellite bus standard to be used in scientific and technology demonstration small missions. Main design drivers are: extended use of microtechnology, nanotechnology and COTS (Commercial-Off-The-Shelf) components, and formation flying capabilities. The first flight is expected by the end of 2008. Aiming at the microsatellite realization, ASI has financed several research projects devoted to the development of enabling technologies, such as micro-sensors and microactuators. In this framework, the research team at the university of Naples is developing the prototype of a two-axis, APS-based micro sun sensor and a laboratory facility for

353

sensor indoor testing and calibration. This paper describes the sun sensor concept, architecture and prototype and the indoor testing facility, as well as results of the sensor validation and calibration performed using neural-network-based techniques.

2. SENSOR CONCEPT The sensor optical head consists of an APS photodetector covered by an opaque mask with one or more tiny holes. The mask is placed at a distance of 3 mm from the focal plane, allowing a field of view of about ±60°. Two different mask configurations have been developed. The first one has a single hole, whilst the second one has an array of equally-spaced 100 holes, aiming at improving sensor measurement precision by averaging multiple sun images [4], Incoming sun light projects sun images onto the focal plane (Pigure 1). Sun line orientation is then determined by means of sun image centroiding [4-6]. In previous papers, the authors have widely described the centroiding technique as applied to sun images acquired with the one-hole mask and the geometrical model relating sun line orientation to centroid coordinates [4-5]. Concerning operation with multiple sun spots, sun line orientation is derived from the average of the coordinates of 100 centroids (average centroid) on the focal plane. To this end, a technique has been developed consisting of two main steps. First of all, a grid is constructed to separate the imaged spots. It is obtained by eliminating pixels with intensity below a prefixed threshold, representing average image noise, and setting separation lines in the χ and y axes, as shown in Fig.2. Figure 3 shows the result of the described procedure.

Fig.2 Spot intensity distribution and separation grid construction

354

Once the separation grid has been set, for each element of the grid the centroiding window is selected. An extended analysis has demonstrated that centroid coordinate uncertainty in sequential sun acquisitions, at the same orientation, is minimized by selecting the window size using a threshold for pixel elimination equal to 10% of the maximum intensity (see Fig.3). Then the centroid of each spot is evaluated using standard expressions:

Σ ν * •