Deep Reinforcement Learning for Spacecraft Proximity Operations Guidance

2021/01/20 by Kirk Hovell, Steve Ulrich · 4 citations
Engineering · Physics and Astronomy · #Space Satellite Systems and Control #Spacecraft Dynamics and Control #Astro and Planetary Science

paper · doi:10.2514/1.a34838

Abstract

This paper introduces a guidance strategy for spacecraft proximity operations, which leverages deep reinforcement learning, a branch of artificial intelligence. This technique enables guidance strategies to be learned rather than designed. The learned guidance strategy feeds velocity commands to a conventional controller to track. Control theory is used alongside deep reinforcement learning to lower the learning burden and facilitate the transfer of the learned behavior from simulation to reality. In this paper, a proof-of-concept spacecraft pose tracking and docking scenario is considered, in simulation and experiment, to test the feasibility of the proposed approach. Results show that such a system can be trained entirely in simulation and transferred to reality with comparable performance.

Cited by

• Generation of Spacecraft Operations Procedures Using Deep Reinforcement Learning
• Learning Adaptive Multi-Task Guidance, Navigation, and Control via Hypernetworks
• Reinforcement Learning based 6-DoF Maneuvers for Microgravity Intravehicular Docking: A Simulation Study with Int-Ball2 in ISS-JEM
• RL-AVIST: Reinforcement Learning for Autonomous Visual Inspection of Space Targets

Related

• Autonomous Six-Degree-of-Freedom Spacecraft Docking Maneuvers via Reinforcement Learning
• Multiphase Autonomous Docking via Model-Based and Hierarchical Reinforcement Learning
• Reinforcement Learning for Docking Maneuvers with Prescribed Performance
• A Survey on Artificial Intelligence Trends in Spacecraft Guidance Dynamics and Control
• Generation of Spacecraft Operations Procedures Using Deep Reinforcement Learning