发表机构
Center for Humanoid Research, Korea Institute of Science and Technology (KIST); School of Electrical Engineering, Korea University（韩国科学技术研究院类人机器人研究中心; 高丽大学电气工程学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
ProxPI以标称值为中心的MPPI采样结合软邻近代价融入策略，解决学习先验不匹配下基于采样的MPPI性能下降问题，在分布内外任务均表现稳健。
AI 中文摘要
将学习到的策略与模型预测控制相结合，可利用学习到的任务先验，同时保留对新目标和约束的在线适应能力，但当策略分布外（out of distribution）时，性能会下降。在策略引导的模型预测路径积分（MPPI）控制中，以策略为中心的热启动方法会将采样分布集中在策略输出上。当先验不匹配时，将采样分布集中在策略输出上会限制对不合适解的探索，阻碍向任务最优解的恢复。我们提出近端先验注入（Proximal Prior Injection，ProxPI），其保留以标称值为中心的MPPI采样，并通过软邻近代价融入策略。该方法在分布内任务上的性能与现有先验注入方案相当，同时使优化器能够摆脱不准确的策略，恢复到普通MPPI的性能水平。我们从理论上证明，每次更新时以先验为中心的重新中心化会丢弃优化器的修正，而以标称值为中心的采样会保留该修正，并收敛到由任务代价和先验共同确定的解集，且该性能缺陷不会因更大的展开预算（rollout budget）而消除。仿真和真实机器人实验表明，ProxPI在分布内和分布外任务下均表现出稳健的性能。
英文摘要
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.
Comments12 pages, 8 figures
