No AccessFull-Length Papers

Deep Learning Method for Martian Atmosphere Reconstruction

University of Colorado, Boulder, Colorado

*Postdoctoral Research Associate, Smead Aerospace Engineering Sciences; currently Lecturer, Department of Aeronautics, Imperial College London, London, England SW7 2AZ, United Kingdom. Member AIAA.

Search for more papers by this author

and

Jay W. McMahon

https://orcid.org/0000-0002-1847-4795

University of Colorado, Boulder, Colorado

†Assistant Professor, Smead Aerospace Engineering Sciences. Fellow AIAA.

Search for more papers by this author

Published Online:31 Aug 2021https://doi.org/10.2514/1.I010922

Abstract

The reconstruction of atmospheric properties encountered during Mars entry trajectories is a crucial element of postflight mission analysis. This paper proposes a deep learning architecture using a long short-term memory (LSTM) network for the reconstruction of Martian density and wind profiles from inertial measurements and guidance commands. The LSTM is trained on a large set of Mars entry trajectories controlled through the fully numerical predictor-corrector entry guidance (FNPEG) algorithm, with density and wind from the Mars Global Reference Atmospheric Model (GRAM) 2010. The training of the network is examined, ensuring that the LSTM generalizes well to samples not present in the training set, and the performance of the network is assessed on a separate training set. The errors of the reconstructed density and wind profiles are, respectively, within 0.54 and 1.9%. Larger wind errors take place at high altitudes due to the decreased sensitivity of the trajectory in regions of low dynamic pressure. The LSTM architecture reliably reproduces the atmospheric density and wind encountered during descent.

References

[1] Spencer D. A., Blanchard R. C., Braun R. D., Kallemeyn P. H. and Thurman S. W., “Mars Pathfinder Entry, Descent, and Landing Reconstruction,” Journal of Spacecraft and Rockets, Vol. 36, No. 3, 1999, pp. 357–366. https://doi.org/10.2514/2.3478 Link Google Scholar
[2] Tolson R., Bemis E., Hough S., Zaleski K., Keating G., Shidner J., Brown S., Brickler A., Scher M. and Thomas P., “Atmospheric Modeling Using Accelerometer Data During Mars Reconnaissance Orbiter Aerobraking Operations,” Journal of Spacecraft and Rockets, Vol. 45, No. 3, 2008, pp. 511–518. https://doi.org/10.2514/1.34301 Link Google Scholar
[3] Desai P. N., “All Recent Mars Landers Have Landed Downrange—Are Mars Atmosphere Models Mis-Predicting Density?” LPI Contributions, Lunar and Planetary Inst., Williamsburg, VA, 2008. Google Scholar
[4] Dutta S., Braun R. D., Russell R. P., Striepe S. A. and Clark I. G., “Comparison of Statistical Estimation Techniques for Mars Entry, Descent, and Landing Reconstruction,” Journal of Spacecraft and Rockets, Vol. 50, No. 6, 2013, pp. 1207–1221. https://doi.org/10.2514/1.A32459 Link Google Scholar
[5] Dutta S. and Braun R., “Mars Entry, Descent, and Landing Trajectory and Atmosphere Reconstruction,” 48th AIAA Aerospace Sciences Meeting, AIAA Paper 2010-1210, 2010. https://doi.org/10.2514/6.2010-1210 Link Google Scholar
[6] Chen A., Cianciolo A., Vasavada A. R., Karlgaard C., Barnes J., Cantor B., Kass D., Rafkin S. and Tyler D., “Reconstruction of Atmospheric Properties from Mars Science Laboratory Entry, Descent, and Landing,” Journal of Spacecraft and Rockets, Vol. 51, No. 4, 2014, pp. 1062–1075. https://doi.org/10.2514/1.A32708 Link Google Scholar
[7] Golombek M. P., Grant J. A., Parker T. J., Kass D. M., Crisp J. A., Squyres S. W., Haldemann A. F. C., Adler M., Lee W. J., Bridges N. T., Arvidson R. E., Carr M. H., Kirk R. L., Knocke P. C., Roncoli R. B., Weitz C. M., Schofield J. T., Zurek R. W., Christensen P. R., Fergason R. L., Anderson F. S. and Rice J. W., “Selection of the Mars Exploration Rover Landing Sites,” Journal of Geophysical Research: Planets, Vol. 108, No. E12, 2003. https://doi.org/10.1029/2003JE002074 Google Scholar
[8] Arridge S., Maass P., Öktem O. and Schönlieb C.-B., “Solving Inverse Problems Using Data-Driven Models,” Acta Numerica, Vol. 28, May 2019, pp. 1–174. https://doi.org/10.1017/S0962492919000059 Crossref Google Scholar
[9] Goodfellow I., Bengio Y. and Courville A., “Deep Learning,” Adaptive Computation and Machine Learning, MIT Press, Cambridge, MA, 2016, http://www.deeplearningbook.org/. Google Scholar
[10] Hornik K., Stinchcombe M. and White H., “Multilayer Feedforward Networks Are Universal Approximators,” Neural Networks, Vol. 2, No. 5, 1989, pp. 359–366. https://doi.org/10.1016/0893-6080(89)90020-8 Crossref Google Scholar
[11] Schäfer A. M. and Zimmermann H. G., “Recurrent Neural Networks Are Universal Approximators,” Artificial Neural Networks ICANN 2006, edited by Kollias S. D., Stafylopatis A., Duch W. and Oja E., Springer, Berlin, 2006, pp. 632–640. https://doi.org/10.1007/11840817_66 Crossref Google Scholar
[12] Wagner J., Wilhite A., Stanley D. and Powell R., “An Adaptive Real Time Atmospheric Prediction Algorithm for Entry Vehicles,” 3rd AIAA Atmospheric Space Environments Conference, AIAA Paper 2011-3200, 2011. https://doi.org/10.2514/6.2011-3200 Google Scholar
[13] Wilkinson M. C. and Meade A. J., “Neural-Network-Inspired Machine Learning for Autonomous Lunar Targeting,” Journal of Aerospace Information Systems, Vol. 11, No. 7, 2014, pp. 458–466. https://doi.org/10.2514/1.I010166 Link Google Scholar
[14] Hochreiter S. and Schmidhuber J., “Long Short-Term Memory,” Neural Computation, Vol. 9, No. 8, 1997, pp. 1735–1780. https://doi.org/10.1162/neco.1997.9.8.1735 Crossref Google Scholar
[15] Justh H. L., “Mars Global Reference Atmospheric Model 2010 Version: Users Guide,” NASA Marshall Spaceflight Center TM-2014-217499, Huntsville, AL, 2014. Google Scholar
[16] Graves A., “Supervised Sequence Labelling with Recurrent Neural Networks,” Studies in Computational Intelligence, Vol. 385, Springer, Berlin, 2012. https://doi.org/10.1007/978-3-642-24797-2 Google Scholar
[17] Alpaydin E., Introduction to Machine Learning, 3rd ed., Adaptive Computation and Machine Learning, MIT Press, Cambridge, MA, 2014. Google Scholar
[18] “Machine Learning Crash Course,” Google, 2020, https://developers.google.com/machine-learning/crash-course. Google Scholar
[19] Sherstinsky A., “Fundamentals of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) Network,” Physica D: Nonlinear Phenomena, Vol. 404, March 2020, Paper 132306. https://doi.org/10.1016/j.physd.2019.132306 Crossref Google Scholar
[20] Montavon G., Orr G. B. and Müller K.-R. (eds.), Neural Networks: Tricks of the Trade, 2nd ed., Lecture Notes in Computer Science, Vol. 7700, Springer, Berlin, 2012. https://doi.org/10.1007/978-3-642-35289-8 Crossref Google Scholar
[21] Lu P., “Entry Guidance: A Unified Method,” Journal of Guidance, Control, and Dynamics, Vol. 37, No. 3, 2014. https://doi.org/10.2514/1.62605 Link Google Scholar
[22] Lu P., Brunner C. W., Stachowiak S. J., Mendeck G., Tigges M. A. and Cerimele C. J., “Verification of a Fully Numerical Entry Guidance Algorithm,” Journal of Guidance, Control, and Dynamics, Vol. 40, No. 2, 2017, pp. 230–247. https://doi.org/10.2514/1.G000327 Link Google Scholar
[23] Dwyer Cianciolo A. M. and Polsgrove T., “Human Mars Entry, Descent, and Landing Architecture Study: Phase 2 Summary,” AIAA Paper 2018-5190, 2018. https://doi.org/10.2514/6.2018-5190 Google Scholar
[24] Bose D. M., Shidner J., Winski R., Zumwalt C., Cheatwood F. M. and Hughes S. J., “The Hypersonic Inflatable Aerodynamic Decelerator (HIAD) Mission Applications Study,” AIAA Aerodynamic Decelerator Systems (ADS) Conference, AIAA Paper 2013-1389, 2013. https://doi.org/10.2514/6.2013-1389 Link Google Scholar
[25] Haberle R. M., The Atmosphere and Climate of Mars, Cambridge Univ. Press, Cambridge, England, U.K., 2017, p. 233. https://doi.org/10.1017/9781139060172 Crossref Google Scholar
[26] Kingma D. P. and Ba J., “Adam: A Method for Stochastic Optimization,” arXiv:1412.6980 [cs], 2017, http://arxiv.org/abs/1412.6980. Google Scholar
[27] Keskar N. S., Mudigere D., Nocedal J., Smelyanskiy M. and Tang P. T. P., “On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima,” arXiv:1609.04836 [cs, math], 2017, http://arxiv.org/abs/1609.04836. Google Scholar
[28] Sharaf O., Amiri S., AlDhafri S., Withnell P. and Brain D., “Sending Hope to Mars,” Nature Astronomy, Vol. 4, No. 7, 2020, Paper 722. https://doi.org/10.1038/s41550-020-1151-y Crossref Google Scholar
[29] Dwyer-Cianciolo A. M., Sostaric R. R., Lugo R. A., Woffinden D., Karlgaard C. and Chen P.-T., “A Simulation Framework for Precision Landing and Hazard Avoidance Technology Assessments,” AIAA Scitech 2020 Forum, AIAA Paper 2020-0366, 2020. https://doi.org/10.2514/6.2020-0366 Link Google Scholar
[30] Masciarelli J., Rousseau S., Fraysse H. and Perot E., “An Analytic Aerocapture Guidance Algorithm for the Mars Sample Return Orbiter,” Atmospheric Flight Mechanics Conference, AIAA Paper 2000-4116, 2000. https://doi.org/10.2514/6.2000-4116 Link Google Scholar
[31] Perot E. and Rousseau S., “Importance of an On-Board Estimation of the Density Scale Height for Various Aerocapture Guidance Algorithms,” AIAA/AAS Astrodynamics Specialist Conference and Exhibit, AIAA Paper 2002-4734, 2002. https://doi.org/10.2514/6.2002-4734 Link Google Scholar
[32] Johnson B. J., Rocca-Bejar D., Lu P., Nikaido B., Hays Z. B., D’Souza S. and Sostaric R. R., “Pterodactyl: Development and Performance of Guidance Algorithms for a Mechanically Deployed Entry Vehicle,” AIAA Scitech 2020 Forum, AIAA Paper 2020-1011, 2020. https://doi.org/10.2514/6.2020-1011 Google Scholar
[33] Cluzeau J. M., Henriquel X., Rebender G., Soudain G., van Dijk L., Gronskiy A., Haber D., Perret-Gentil C. and Polak R., “Concepts of Design Assurance for Neural Networks (CoDANN),” European Union Aviation Safety Agency, TR, 2020, https://www.easa.europa.eu/document-library/general-publications/concepts-design-assurance-neural-networks-codann. Google Scholar

Journal of Aerospace Information Systems, volume 18 issue 10 cover

Volume 18, Number 10October 2021

Metrics

Crossmark

Information

Copyright © 2021 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved. All requests for copying and permission to reprint should be submitted to CCC at www.copyright.com; employ the eISSN 2327-3097 to initiate your request. See also AIAA Rights and Permissions www.aiaa.org/randp.

Acknowledgments

This work was carried out under the NASA Early Career Fellowship (ECF) grant 80NSSC18K1510 on “Robust Entry and Landing Guidance Under Dynamic Uncertainty.” Davide Amato thanks Evan Roelke (University of Colorado Boulder) for providing information on the Hypersonic Inflatable Aerodynamic Decelerator vehicle, Andrew Harris (University of Colorado Boulder) for helpful discussions on LSTM training, and Roberto Furfaro (The University of Arizona) for having introduced him to neural networks.

PDF

Received16 October 2020
Accepted14 July 2021
Published online31 August 2021

Deep Learning Method for Martian Atmosphere Reconstruction

Abstract

References

What's Popular

Metrics

Information