Distributionally Robust Adaptive Iterative Covariance Steering for Small-Body Proximity Operations
We combine iterative covariance steering (iCS) with $L1$ distributionally robust adaptive control ($L1$-DRAC) to provide risk-aware collision avoidance despite mismatch between the nominal planning model and the true dynamics.
ProofPaper ↗
Key points
- Autonomous small-body proximity operations require continuous-thrust guidance under uncertain gravity, solar-radiation pressure (SRP), and stochastic disturbances.
- Iterative covariance steering plans a nominal feedback policy satisfying distributionally robust conditional value-at-risk keep-out constraints, while $L1$-DRAC compensates for dynamics uncertainty and bounds the residual distributional mismatch through a Wasserstein radius used by the planner.
- In $1000$ paired Monte Carlo trials with gravity, SRP, and diffusion mismatch, $L1$-DRAC adaptive control reduces the keep-out violation rate from $32.7%$ to $0.2%$ and the physical collision rate from $16.4%$ to zero, without actuator saturation.
Sources (1)
- [1]Distributionally Robust Adaptive Iterative Covariance Steering for Small-Body Proximity OperationsarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 10:27 PM
We combine iterative covariance steering (iCS) with $L_1$ distributionally robust adaptive control ($L_1$-DRAC) to provide risk-aware collision avoidance despite mismatch between the nominal planning model and the true dynamics.
Autonomous small-body proximity operations require continuous-thrust guidance under uncertain gravity, solar-radiation pressure (SRP), and stochastic disturbances.
Extractive summary: sentences quoted from the sources.