Adversarial learning improves robustness of spacecraft landing systems
Adversarial Reinforcement Learning for Robust Planetary Landing
Machine Learning
Summary
Spacecraft landing systems often face tough conditions that are hard to predict because disturbances during landing can happen in complex, related ways rather than as simple random events. The authors show that using an approach where an artificial opponent deliberately creates challenging disturbances helps reveal weaknesses in landing controllers that usual random testing misses. By training two agents—one controlling the landing and one creating tough disturbances—the landing control improves its ability to handle difficult and unexpected situations. This method hints that spacecraft could land more safely even under unusual landing disturbances.
What this means in practice
- •For aerospace control engineers: Enhance spacecraft landing controllers by training with adversarial disturbances to improve performance under complex, correlated real-world landing challenges.
- •For robotics system developers: Use adversarial reinforcement learning to harden robot control policies against challenging correlated disturbances not easily captured by random noise testing.
Authors
Miguel Leiva-Vélez, Nicolas Gaston-Rozado, Andrea Scorsoglio, Hodei Urrutxua, Víctor Rodríguez-Fernández
Abstract
Deep reinforcement learning (DRL) policies for planetary landing and other aerospace guidance, navigation, and control tasks are commonly trained under randomized disturbances to improve robustness. However, typical stochastic perturbation sampling assumes independent variations, whereas real mission disturbances often exhibit spatial and temporal correlation and cannot be reproduced through independent random draws. This work investigates adversarial reinforcement learning to discover challenging off-nominal conditions. A secondary agent generates physically constrained perturbations that degrade the performance of a landing controller, forming an asymmetric sequential zero-sum game. Training proceeds in three stages: baseline controller learning, robustness enhancement with stochastic perturbations, and an iterative adversarial phase in which the controller and disturbance generator are alternately optimized. Preliminary experiments in a three-degree-of-freedom landing environment show that adversarial disturbances reveal vulnerabilities missed by conventional randomization and provide early indications that, with a more extensive training campaign, the approach can yield improved robustness and stronger generalization under previously unseen perturbation regimes.