A New Method Based on a Multilayer Perceptron Network to Determine In-Orbit Satellite Attitude for Spacecrafts without Active ADCS Like UVSQ-SAT
<p>General method to determine the terrestrial net radiation from UVSQ-SAT remote sensors.</p> "> Figure 2
<p>Ground-based calibration implemented to train and validate the satellite attitude determination algorithm.</p> "> Figure 3
<p>Representation of the reference basis in the satellite frame of reference.</p> "> Figure 4
<p>Neural Network architecture with fully connected layers, as well as the inputs and outputs.</p> "> Figure 5
<p>Loss functions for the validation set and the train set at each iteration (Epoch) during training.</p> "> Figure 6
<p>Predicted azimuth of the Sun versus real azimuth of reference in degrees based on the test set done in October 2020 and the optimized neural network. The continuous blue curve represents the ideal case when the observed values perfectly fit the predicted values.</p> "> Figure 7
<p>Comparison of the different solar flux computations for the test conducted on 26 October 2020. (<b>A</b>) pyranometer flux corrected of the Sun’s elevation; (<b>B</b>) photodiodes’s flux corrected of the reference attitude; (<b>C</b>) photodiodes’s normalized flux corrected of the predicted attitude.</p> "> Figure 8
<p>Ratio of the flux for the pyranometer and the corrected flux thanks to the reference and predicted attitude. (<b>A</b>) ratio between the reference flux and the predicted flux; (<b>B</b>) ratio between pyranometer flux and the predicted flux; (<b>C</b>) ratio between the pyranometer flux and the reference flux.</p> "> Figure 9
<p>Solar cells temperature for all UVSQ-SAT faces. Measurements start from 7 February 2021 at 12:00 AM (UTC). The orbital period is about 95.18 mn. UVSQ-SAT spins during the orbit.</p> "> Figure 10
<p>Uncalibrated LED flux (photodiodes in the 400–1100 nm wavelength range) for all faces of the UVSQ-SAT CubeSat. Measurements (TSI + OSR) start from 7 February 2021 at 12:00 AM (UTC).</p> "> Figure 11
<p>Magnetic field components on the <span class="html-italic">x</span>-axis, <span class="html-italic">y</span>-axis, and <span class="html-italic">z</span>-axis.</p> ">
Abstract
:1. Introduction
2. Method to Retrieve Terrestrial Net Radiation for Satellite That Does Not Have Active ADCS
- 12 ERS sensors are part of the UVSQ-SAT satellite (2 per face). Six ERS sensors aims to measure radiation between 0.2 and 100 m using Carbon NanoTubes coatings (absorptivity close to 1). Six other ERS sensors aim to measure radiation wavelength between 0.2 and 3 m with 0.06 absorptivity and between 3 and 100 m with 0.84 absorptivity using optical solar reflector coatings. ERS measurements represent indicators to detect Earth and the Sun positions.
- Three ultraviolet sensors (UVS) are part of the UVSQ-SAT scientific payload. They focus on the 200–1100 nm wavelength range.
- Six photodiodes (LED) are located on the spacecraft and measure solar and outgoing shortwave radiations in the 400–1100 nm wavelength range.
- Temperature sensors (solar cells) are also located on each satellite panel.
- Teach’ Wear (TW) is a new three axis accelerometer/gyroscope/compass. The TW module on-board UVSQ-SAT has an instrumentation that will be very helpful to determine the reference position during the training phase such as: a three-axis accelerometer, a three-axis gyrometer, and a three-axis magnetometer.
2.1. General Method Description to Determine the Terrestrial Net Radiation
2.2. Method Based on a Deep Learning Approach to Determine Satellite Attitude
2.2.1. Training of the Deep Learning Neural Network
2.2.2. Neural Network Architecture of the Deep Learning Method
- 5 Hidden fully connected layers
- 25 Inputs
- 2 Outputs
- Learning rate of , determined empirically
- Layers dimensions (width): 25/48/128/256/128/2
2.2.3. Loss Function for Training the Deep Learning Neural Network
2.2.4. Performance and Uncertainties
3. Results
4. Discussion and Perspectives
5. Conclusions
Author Contributions
Funding
Acknowledgments
Conflicts of Interest
References
- Meftah, M.; Damé, L.; Keckhut, P.; Bekki, S.; Sarkissian, A.; Hauchecorne, A.; Bertran, E.; Carta, J.P.; Rogers, D.; Abbaki, S.; et al. UVSQ-SAT, a pathfinder cubesat mission for observing essential climate variables. Remote Sens. 2020, 12, 92. [Google Scholar] [CrossRef] [Green Version]
- Meftah, M.; Damé, L.; Bolsée, D.; Hauchecorne, A.; Pereira, N.; Sluse, D.; Cessateur, G.; Irbah, A.; Bureau, J.; Weber, M.; et al. SOLAR-ISS: A new reference spectrum based on SOLAR/SOLSPEC observations. Astron. Astrophys. 2018, 611, A1. [Google Scholar] [CrossRef]
- Meftah, M.; Snow, M.; Damé, L.; Bolseé, D.; Pereira, N.; Cessateur, G.; Bekki, S.; Keckhut, P.; Sarkissian, A.; Hauchecorne, A. SOLAR-v: A new solar spectral irradiance dataset based on SOLAR/SOLSPEC observations during solar cycle 24. Astron. Astrophys. 2021, 2, 1–9. [Google Scholar] [CrossRef]
- Stephens, G.L.; Li, J.; Wild, M.; Clayson, C.A.; Loeb, N.; Kato, S.; L’Ecuyer, T.; Stackhouse, P.W.; Lebsock, M.; Andrews, T. An update on Earth’s energy balance in light of the latest global observations. Nat. Geosci. 2012, 5, 691–696. [Google Scholar] [CrossRef]
- Allan, R.P.; Liu, C.; Loeb, N.G.; Palmer, M.D.; Roberts, M.; Smith, D.; Vidale, P.L. Changes in global net radiative imbalance 1985–2012. Geophys. Res. Lett. 2014, 41, 5588–5597. [Google Scholar] [CrossRef] [PubMed]
- Wild, M.; Ohmura, A.; Schär, C.; Müller, G.; Folini, D.; Schwarz, M.; Hakuba, M.Z.; Sanchez-Lorenzo, A. The Global Energy Balance Archive (GEBA) version 2017: A database for worldwide measured surface energy fluxes. Earth Syst. Sci. Data 2017, 9, 601–613. [Google Scholar] [CrossRef] [Green Version]
- Brasoveanu, D.; Hashmall, J.; Baker, D. Spacecraft Attitude Determination Accuracy from Mission Experience. NASA GSFC Tech. Rep. 1994, 95, 12858. [Google Scholar]
- Markley, F.L.; Crassidis, J.L. Static Attitude Determination Methods. In Fundamentals of Spacecraft Attitude Determination and Control; Springer: New York, NY, USA, 2014; pp. 183–233. [Google Scholar] [CrossRef]
- Iwasaki, Y.; Kikuya, Y.; Sasaki, K.; Ozawa, T.; Shintani, Y.; Masuda, Y.; Watanabe, K.; Mamiya, H.; Ando, H.; Nakashima, T.; et al. Development and Initial On-orbit Performance of Multi-Functional Attitude Sensor using Image Recognition. In Proceedings of the AIAA/USU Conference on Small Satellites, Logan, UT, USA, 3–8 August 2019. [Google Scholar]
- Zhu, X.X.; Tuia, D.; Mou, L.; Xia, G.; Zhang, L.; Xu, F.; Fraundorfer, F. Deep Learning in Remote Sensing: A Comprehensive Review and List of Resources. IEEE Geosci. Remote Sens. Mag. 2017, 5, 8–36. [Google Scholar] [CrossRef] [Green Version]
- Malerba, M.E.; Wright, N.; Macreadie, P.I. A continental-scale assessment of density, size, distribution and historical trends of farm dams using deep learning convolutional neural networks. Remote Sens. 2021, 13, 319. [Google Scholar] [CrossRef]
- Debella-Gilo, M.; Gjertsen, A.K. Mapping Seasonal Agricultural Land Use Types Using Deep Learning on Sentinel-2 Image Time Series. Remote Sens. 2021, 13, 289. [Google Scholar] [CrossRef]
- Rostami, M.; Kolouri, S.; Eaton, E.; Kim, K. Deep transfer learning for few-shot SAR image classification. Remote Sens. 2019, 11, 1374. [Google Scholar] [CrossRef] [Green Version]
- Shao, Z.; Zhou, Z.; Huang, X.; Zhang, Y. Mrenet: Simultaneous extraction of road surface and road centerline in complex urban scenes from very high-resolution images. Remote Sens. 2021, 13, 239. [Google Scholar] [CrossRef]
- Reda, I.; Nrel, A.A. Solar Position Algorithm for Solar Radiation Applications (Revised); National Renewable Energy Laboratory Nrel/Tp-560-34302; National Renewable Energy Laboratory: Golden, CO, USA, 2005; pp. 1–56. [Google Scholar]
- Finance, A.; Meftah, M.; Mangin, A.; Keckhut, P. UVSQ-SAT a New Way to Obtain Spatio-Temporal Variations of the Radiation Budget with a Satellite Constellation; Earth and Space Science Open Archive; American Geophysical Union: Washington, DC, USA, 2021; p. 21. [Google Scholar] [CrossRef]
- Nie, F.; Hu, Z.; Li, X. An investigation for loss functions widely used in machine learning. Commun. Inf. Syst. 2018, 18, 37–52. [Google Scholar] [CrossRef]
- Martel, F.; Pal, P.K.; Psiaki, M.L. Three-axis attitude determination via Kalman filtering of magnetometer data. J. Guid. Control. Dyn. 1990, 13, 344–367. [Google Scholar]
Activation Function | Mean Squared Error | Iteration |
---|---|---|
RELU | 55.6 | 282 |
ELU | 61.8 | 284 |
Tanh | 61.1 | 288 |
Sigmoid | 6259.5 | 240 |
Leaky RELU | 59.2 | 260 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Finance, A.; Meftah, M.; Dufour, C.; Boutéraon, T.; Bekki, S.; Hauchecorne, A.; Keckhut, P.; Sarkissian, A.; Damé, L.; Mangin, A. A New Method Based on a Multilayer Perceptron Network to Determine In-Orbit Satellite Attitude for Spacecrafts without Active ADCS Like UVSQ-SAT. Remote Sens. 2021, 13, 1185. https://doi.org/10.3390/rs13061185
Finance A, Meftah M, Dufour C, Boutéraon T, Bekki S, Hauchecorne A, Keckhut P, Sarkissian A, Damé L, Mangin A. A New Method Based on a Multilayer Perceptron Network to Determine In-Orbit Satellite Attitude for Spacecrafts without Active ADCS Like UVSQ-SAT. Remote Sensing. 2021; 13(6):1185. https://doi.org/10.3390/rs13061185
Chicago/Turabian StyleFinance, Adrien, Mustapha Meftah, Christophe Dufour, Thomas Boutéraon, Slimane Bekki, Alain Hauchecorne, Philippe Keckhut, Alain Sarkissian, Luc Damé, and Antoine Mangin. 2021. "A New Method Based on a Multilayer Perceptron Network to Determine In-Orbit Satellite Attitude for Spacecrafts without Active ADCS Like UVSQ-SAT" Remote Sensing 13, no. 6: 1185. https://doi.org/10.3390/rs13061185