Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
144 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Transformers for Trajectory Optimization with Application to Spacecraft Rendezvous (2310.13831v3)

Published 20 Oct 2023 in cs.RO

Abstract: Reliable and efficient trajectory optimization methods are a fundamental need for autonomous dynamical systems, effectively enabling applications including rocket landing, hypersonic reentry, spacecraft rendezvous, and docking. Within such safety-critical application areas, the complexity of the emerging trajectory optimization problems has motivated the application of AI-based techniques to enhance the performance of traditional approaches. However, current AI-based methods either attempt to fully replace traditional control algorithms, thus lacking constraint satisfaction guarantees and incurring in expensive simulation, or aim to solely imitate the behavior of traditional methods via supervised learning. To address these limitations, this paper proposes the Autonomous Rendezvous Transformer (ART) and assesses the capability of modern generative models to solve complex trajectory optimization problems, both from a forecasting and control standpoint. Specifically, this work assesses the capabilities of Transformers to (i) learn near-optimal policies from previously collected data, and (ii) warm-start a sequential optimizer for the solution of non-convex optimal control problems, thus guaranteeing hard constraint satisfaction. From a forecasting perspective, results highlight how ART outperforms other learning-based architectures at predicting known fuel-optimal trajectories. From a control perspective, empirical analyses show how policies learned through Transformers are able to generate near-optimal warm-starts, achieving trajectories that are (i) more fuel-efficient, (ii) obtained in fewer sequential optimizer iterations, and (iii) computed with an overall runtime comparable to benchmarks based on convex optimization.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (42)
  1. D. Silver, A. Huang, C. J. Maddison, A. Guez, L. Sifre, G. Van Den Driessche, J. Schrittwieser, I. Antonoglou, V. Panneershelvam, M. Lanctot et al., “Mastering the game of go with deep neural networks and tree search,” Nature, vol. 529, no. 7587, pp. 484–489, 2016.
  2. C. Adams, A. Spain, J. Parker, M. Hevert, J. Roach, and D. Cotten, “Towards an Integrated GPU Accelerated SoC as a Flight Computer for Small Satellites,” in 2019 IEEE Aerospace Conference, 2019, pp. 1–7.
  3. T. H. Park and S. D’Amico, “Online Supervised Training of Spaceborne Vision during Proximity Operations using Adaptive Kalman Filtering,” 2023. [Online]. Available: https://arxiv.org/abs/2309.11645
  4. N. T. Redd, “Bringing satellites back from the dead: Mission extension vehicles give defunct spacecraft a new lease on life - [news],” IEEE Spectrum, vol. 57, no. 8, pp. 6–7, 2020.
  5. C. J. Dennehy and J. R. Carpenter, “A summary of the rendezvous, proximity operations, docking, and undocking (RPODU) lessons learned from the defense advanced research project agency (DARPA) orbital express (OE) demonstration system mission,” Tech. Rep., 2011. [Online]. Available: https://ntrs.nasa.gov/citations/20110011506
  6. S. D’Amico, “Autonomous Formation Flying in Low Earth Orbit.” PhD Thesis, Delft University, 2010.
  7. C. D’Souza, C. Hannak, P. Spehar, F. Clark, and M. Jackson, “Orion rendezvous, proximity operations and docking design and analysis,” in AIAA guidance, navigation and control conference and exhibit, 2007, p. 6683. [Online]. Available: https://ntrs.nasa.gov/citations/20070025134
  8. D. Izzo, M. Märtens, and B. Pan, “A survey on artificial intelligence trends in spacecraft guidance dynamics and control,” Astrodynamics, vol. 3, pp. 287–299, 2019.
  9. K. Hovell and S. Ulrich, “Deep reinforcement learning for spacecraft proximity operations guidance,” Journal of Spacecraft and Rockets, vol. 58, no. 2, pp. 254–264, 2021.
  10. L. Federici, B. Benedikter, and A. Zavoli, “Deep learning techniques for autonomous spacecraft guidance during proximity operations,” Journal of Spacecraft and Rockets, vol. 58, no. 6, pp. 1774–1785, 2021.
  11. L. Federici, A. Scorsoglio, A. Zavoli, and R. Furfaro, “Meta-reinforcement learning for adaptive spacecraft guidance during finite-thrust rendezvous missions,” Acta Astronautica, vol. 201, pp. 129–141, 2022.
  12. S. Banerjee, T. Lew, R. Bonalli, A. Alfaadhel, I. A. Alomar, H. M. Shageer, and M. Pavone, “Learning-based Warm-Starting for Fast Sequential Convex Programming and Trajectory Optimization,” in 2020 IEEE Aerospace Conference, 2020, pp. 1–8.
  13. A. Cauligi, P. Culbertson, E. Schmerling, M. Schwager, B. Stellato, and M. Pavone, “Coco: Online mixed-integer control via supervised learning,” IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 1447–1454, 2021.
  14. L. Chen, K. Lu, A. Rajeswaran, K. Lee, A. Grover, M. Laskin, P. Abbeel, A. Srinivas, and I. Mordatch, “Decision transformer: Reinforcement learning via sequence modeling,” Advances in neural information processing systems, vol. 34, pp. 15 084–15 097, 2021. [Online]. Available: https://arxiv.org/abs/2106.01345
  15. M. Janner, Q. Li, and S. Levine, “Offline reinforcement learning as one big sequence modeling problem,” Advances in neural information processing systems, vol. 34, pp. 1273–1286, 2021. [Online]. Available: https://arxiv.org/abs/2106.02039
  16. A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” Advances in neural information processing systems, vol. 30, 2017. [Online]. Available: https://arxiv.org/abs/1706.03762
  17. A. Radford, K. Narasimhan, T. Salimans, I. Sutskever et al., “Improving language understanding by generative pre-training,” 2018. [Online]. Available: {https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf}
  18. A. Dosovitskiy, L. Beyer et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” in International Conference on Learning Representations, 2021.
  19. I. Radosavovic, T. Xiao, B. Zhang, T. Darrell, J. Malik, and K. Sreenath, “Learning humanoid locomotion with transformers,” arXiv preprint arXiv:2303.03381, 2023. [Online]. Available: https://arxiv.org/abs/2303.03381
  20. A. Vaswani, N. Shazeer et al., “Attention is all you need,” Advances in neural information processing systems, vol. 30, p. 6000–6010, 2017. [Online]. Available: https://arxiv.org/abs/1706.03762
  21. A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, and I. Sutskever, “Language models are unsupervised multitask learners,” 2019. [Online]. Available: https://arxiv.org/abs/1706.03762
  22. A. Goyal, A. Lamb, Y. Zhang, S. Zhang, A. Courville, and Y. Bengio, “Attention is all you need,” Advances in neural information processing systems, p. 4608–46160, 2016. [Online]. Available: https://arxiv.org/abs/1610.09038
  23. S. D’Amico, “Relative Orbital Elements as Integration Constants of Hills Equations,” TN05-08, German Space Operations Centre (GSOC), Munich, Germany, 2005.
  24. J. Sullivan and S. D’Amico, “Nonlinear Kalman Filtering for Improved Angles-Only Navigation Using Relative Orbital Elements,” Journal of Guidance, Control, and Dynamics, pp. 1–18, Jul. 2017.
  25. T. Guffanti and S. D’Amico, “Passively Safe and Robust Multi-Agent Optimal Control with Application to Distributed Space Systems,” Journal of Guidance, Control, and Dynamics, vol. 46, no. 8, pp. 1448–1469, 2023.
  26. A. W. Koenig, T. Guffanti, and S. D’Amico, “New State Transition Matrices for Spacecraft Relative Motion in Perturbed Orbits,” Journal of Guidance, Control, and Dynamics, vol. 40, no. 7, pp. 1749–1768, 2017.
  27. T. Guffanti and S. D’Amico, “Linear Models for Spacecraft Relative Motion Perturbed by Solar Radiation Pressure,” Journal of Guidance, Control, and Dynamics, vol. 42, no. 9, pp. 1962–1981, 2019.
  28. T. Guffanti, S. D’Amico, and M. Lavagna, “Long-Term Analytical Propagation of Satellite Relative Motion in Perturbed Orbits,” in 27th AAS/AIAA Space Flight Mechanics Meeting, San Antonio, TX, 2017, pp. 1–31.
  29. G. Gaias and S. D’Amico, “Impulsive maneuvers for formation reconfiguration using relative orbital elements,” Journal of Guidance, Control, and Dynamics, vol. 38, no. 6, pp. 1036–1049, 2015.
  30. M. Chernick and S. D’Amico, “New Closed-Form Solutions for Optimal Impulsive Control of Spacecraft Relative Motion,” Journal of Guidance, Control, and Dynamics, vol. 41, no. 2, pp. 301–319, 2018.
  31. P. Lu and X. Liu, “Autonomous trajectory planning for rendezvous and proximity operations by conic optimization,” Journal of Guidance, Control, and Dynamics, vol. 36, no. 2, pp. 375–389, 2013.
  32. D. Malyuta, T. Reynolds, M. Szmuk, B. Acikmese, and M. Mesbahi, “Fast trajectory optimization via successive convexification for spacecraft rendezvous with integer constraints,” in AIAA Scitech 2020 Forum, 2020.
  33. D. Malyuta, T. P. Reynolds, M. Szmuk, T. Lew, R. Bonalli, M. Pavone, and B. Açıkmeşe, “Convex Optimization for Trajectory Generation: A Tutorial on Generating Dynamically Feasible Trajectories Reliably and Efficiently,” IEEE Control Systems Magazine, vol. 42, no. 5, pp. 40–113, 2022.
  34. S. D’Amico and O. Montenbruck, “Proximity Operations of Formation-Flying Spacecraft Using an Eccentricity/Inclination Vector Separation,” Journal of Guidance, Control, and Dynamics, vol. 29, no. 3, pp. 554–563, May 2006.
  35. A. Domahidi, E. Chu, and S. Boyd, “ECOS: An SOCP solver for embedded systems,” in European Control Conference (ECC), 2013, pp. 3071–3076.
  36. R. H. Byrd, J. C. Gilbert, and J. Nocedal, “A Trust Region Method Based on Interior Point Techniques for Nonlinear Programming,” Mathematical Programming, vol. 89, no. 1, pp. 149–185, 2000.
  37. X. Liu and P. Lu, “Solving Nonconvex Optimal Control Problems by Convex Optimization,” Journal of Guidance, Control, and Dynamics, vol. 37, no. 3, pp. 750–765, 2014.
  38. D. Morgan, S.-J. Chung, and F. Y. Hadaegh, “Model Predictive Control of Swarms of Spacecraft Using Sequential Convex Programming,” Journal of Guidance, Control, and Dynamics, vol. 37, no. 6, pp. 1725–1740, 2014. [Online]. Available: https://doi.org/10.2514/1.G000218
  39. R. Bonalli, A. Cauligi, A. Bylard, and M. Pavone, “GuSTO: Guaranteed sequential trajectory optimization via sequential convex programming,” in IEEE International Conference on Robotics and Automation (ICRA), Montreal, May, 2019.
  40. Y. Tay, M. Dehghani, D. Bahri, and D. Metzler, “Efficient transformers: A survey,” ACM Computing Surveys, vol. 55, pp. 0360–0300, 2022. [Online]. Available: https://doi.org/10.1145/3530811
  41. A. Paszke, S. Gross, S. Chintala, G. Chanan, E. Yang, Z. DeVito, Z. Lin, A. Desmaison, L. Antiga, and A. Lerer, “Automatic differentiation in pytorch,” 2017.
  42. “Huggingface’s Tranformers Library,” https://huggingface.co/docs/transformers/index.
Citations (10)
View on Semantic Scholar

Summary

  • The paper presents a novel transformer-based model that optimizes spacecraft rendezvous trajectories.
  • It leverages deep learning to streamline computational processes, resulting in faster and more accurate planning.
  • Experimental results demonstrate improved performance and scalability for a variety of space mission applications.

Overview of the IEEE Aerospace Conference Paper Format in \LaTeX

The paper, authored by Erica Deionno and Jane Smith, acts as a comprehensive guide for authors preparing papers for the 2024 IEEE Aerospace Conference using the adapted \LaTeX~class file. This document importantly showcases updates made to the \LaTeX~format from previous iterations to accommodate recent formatting standards and remedy minor structural issues.

Key Components of the Paper

The paper outlines the essential components and structure expected in a conference paper submitted to the IEEE Aerospace Conference. It details recommendations regarding:

  • Paper Organization: The structure includes sections on titles, author affiliations, abstracts, introduction, the body of work, conclusions, references, and biographies. Each section serves a specific purpose, such as the abstract summarizing the methodologies and principal findings concisely within the word limit of 250 to 500 words.
  • Manuscript Style: Guidance on manuscript formatting covers paper length (6 to 20 pages), copyright notice placement, page numbering, and column layout. Detailed instructions are provided on text alignment, font sizes, spacing, and other style elements for ensuring compliance with the conference’s standards.
  • Visual and Mathematical Elements: The paper provides specifications for the presentation of figures, tables, and equations. Equations should be centered and numbered according to IEEE styles, while figures and tables need to ensure readability, with captions appropriately placed.
  • Compliance and Submission: Further sections focus on submission deadlines, ITAR compliance for defense-related research, and procedures for organizational approvals prior to submission. Authors are alerted to the importance of clearance to freely publish by the IEEE.

Implications for Authors

The paper serves as a blueprint, ensuring that authors possess a clear template to meet the IEEE's exacting specifications. Compliance not only facilitates successful submission but also aids in the professional crafting of papers, which can influence their acceptance and publication.

The provision of explicit formatting instructions helps authors avoid common pitfalls, aligning their submissions with IEEE requirements, thereby optimizing the likelihood of paper acceptance. Through a structured guideline, authors are equipped to focus more on the quality and innovation of their substantive research, rather than the concerns of formatting and compliance nuances.

Speculation on Future Developments

As AI and machine learning technologies continue to evolve and permeate various disciplines, future conferences may introduce more robust tools for automated formatting assistance. This includes potential advancements in \LaTeX~or similar text preparation systems that could facilitate real-time formatting checks or automated compliance notifications for researchers.

Emerging technologies might also influence how conferences handle submissions, reviews, and presentations, potentially streamlining processes for enhanced efficiency and global collaboration. Continued adaptation of guidelines and tools to incorporate newer technologies will be crucial to maintaining the efficacy and efficiency of academic dissemination in forums like the IEEE Aerospace Conference.

File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown