RAT iLQR: A Risk Auto-Tuning Controller to Optimally Account for Stochastic Model Mismatch
This work addresses robotic operation in uncertain environments, offering a method to handle distributional mismatches, but it appears incremental as it builds on existing distributionally robust and risk-sensitive control frameworks.
The paper tackles the problem of robotic control in stochastic environments with imperfect probability distributions by proposing a nonlinear MPC for distributionally robust control, which plans feedback policies against worst-case distributions within a KL divergence bound and dynamically adjusts risk-sensitivity online, demonstrating benefits in a dynamic collision avoidance scenario with erroneous human motion predictions.
Successful robotic operation in stochastic environments relies on accurate characterization of the underlying probability distributions, yet this is often imperfect due to limited knowledge. This work presents a control algorithm that is capable of handling such distributional mismatches. Specifically, we propose a novel nonlinear MPC for distributionally robust control, which plans locally optimal feedback policies against a worst-case distribution within a given KL divergence bound from a Gaussian distribution. Leveraging mathematical equivalence between distributionally robust control and risk-sensitive optimal control, our framework also provides an algorithm to dynamically adjust the risk-sensitivity level online for risk-sensitive control. The benefits of the distributional robustness as well as the automatic risk-sensitivity adjustment are demonstrated in a dynamic collision avoidance scenario where the predictive distribution of human motion is erroneous.