Adaptive Guidance and Integrated Navigation with Reinforcement Meta-Learning
This work addresses the problem of real-time adaptive guidance for autonomous spacecraft landing in highly uncertain environments, offering a method that integrates guidance and navigation using minimal sensor data.
The paper develops an adaptive guidance system using reinforcement meta-learning with recurrent policies, achieving real-time adaptation to unknown dynamics in challenging environments like Mars and asteroid landings. The recurrent policy outperforms non-recurrent RL and traditional guidance laws in tasks with random engine failure and unknown environmental dynamics.
This paper proposes a novel adaptive guidance system developed using reinforcement meta-learning with a recurrent policy and value function approximator. The use of recurrent network layers allows the deployed policy to adapt real time to environmental forces acting on the agent. We compare the performance of the DR/DV guidance law, an RL agent with a non-recurrent policy, and an RL agent with a recurrent policy in four challenging environments with unknown but highly variable dynamics. These tasks include a safe Mars landing with random engine failure and a landing on an asteroid with unknown environmental dynamics. We also demonstrate the ability of a RL meta-learning optimized policy to implement a guidance law using observations consisting of only Doppler radar altimeter readings in a Mars landing environment, and LIDAR altimeter readings in an asteroid landing environment, thus integrating guidance and navigation.