SOURCE-LINKED INTELLIGENCE

ProxPI: Proximal Prior Injection for Sampling-Based MPC under Learned-Prior Mismatch

arXiv · AI, language, vision and robotics · article · Sep 1, 2026 · UTC

Combining learned policies with model predictive control can leverage learned task priors while retaining online adaptation to new objectives and constraints, but performance degrades when the policy is out of distribution. In policy-guided model predictive path integral (MPPI) control, a policy-centered warm-start approach centers the sampling distribution on the policy output. When the prior is mismatched, centering the sampling distribution on the policy output restricts exploration around an unsuitable solution and prevents recovery toward the task optimum. We propose Proximal Prior Inject

Read original source ↗ Open in workspace

recordType

paper

region

Global