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

2026-09-01

AI Summary

本文提出ProxPI方法解决在学习先验不匹配情况下基于采样的MPC性能下降问题，通过软接近成本结合策略保留了探索能力。

Abstract

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 Injection (ProxPI), which retains nominal-centered MPPI sampling and incorporates the policy through a soft proximity cost. This matches the in-distribution performance of existing prior-injection schemes while enabling the optimizer to escape an inaccurate policy and recover vanilla MPPI-level performance. We theoretically show that re-centering on the prior discards the optimizer's correction at every update, whereas nominal-centered sampling retains it and converges to a solution set by both the task cost and the prior, and that this failure is not removed by a larger rollout budget. Simulations and real-robot experiments demonstrate robust performance under both in-distribution and out-of-distribution tasks.

Problem

• Learned Policy
• Model Predictive Control
• Sampling Distribution
• Prior Mismatch
• Performance Degradation

Innovation

• Proximal Prior Injection
• Model Predictive Control
• Sampling Distribution
• Soft Proximity Cost
• Out-of-Distribution