arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2610.38657cs.ROcs.AIcs.LG
收藏

探索关键区域:面向高效漫游车自适应的定向探索

Explore Where It Matters: Targeted Exploration for Efficient Rover Adaptation

University of Texas at Austin(德克萨斯大学奥斯汀分校);Jet Propulsion Laboratory, California Institute of Technology(加州理工学院喷气推进实验室)

Jacob Levy, Arthur Bouton, Tristan D. Hassler, Hari Nayar, David Fridovich-Keil

AI总结:

该研究针对行星漫游车在可变形地形中面临的仿真到真实差距问题,提出TERRA在线自适应框架,通过奖励加权损失和损失预测器引导探索,在仿真和真实试验中均实现了优异的爬坡自适应性能。

摘要:

未来行星漫游车将在陡峭、可变形的地形中行驶,滑动、沉陷和地形变形可能威胁其移动能力。尽管仿真可以捕捉部分轮-土相互作用,但未建模的可变形地形现象会产生仿真到真实的差距,需要部署时的自适应。然而,漫游车操作几乎没有空间进行大量试错。我们提出TERRA(面向奖励相关自适应的定向探索,Targeted Exploration for Reward-Relevant Adaptation),这是一种在线自适应框架,可高效寻找暴露任务相关模型误差的部署数据并进行修正。TERRA通过奖励加权预测损失调整仿真训练的世界模型,并训练损失预测器估计未来与性能相关的误差可能发生的位置,利用该信号在模型预测控制(MPC)期间引导探索。在仿真中,TERRA在分布外预训练的严苛条件下,五次试验内达到约90%的爬坡成功率,在奖励、完成时间和首次成功能耗方面优于探索基线。在ERNEST漫游车上,TERRA在3至5次真实世界试验内恢复了可变形火星模拟物斜坡上的协调主动悬挂移动能力。项目网站:this https URL

英文摘要:

Future planetary rovers will traverse steep, deformable terrain where slip, sinkage, and terrain deformation can threaten mobility. Although simulation can capture some wheel-soil effects, unmodeled deformable-terrain phenomena create sim-to-real gaps requiring deployment-time adaptation. Yet, rover operations leave little room for extensive trial-and-error. We introduce TERRA (Targeted Exploration for Reward-Relevant Adaptation), an online adaptation framework that efficiently seeks deployment data exposing task-relevant model errors and corrects them. TERRA adapts a simulation-trained world model with a reward-weighted prediction loss and trains a loss predictor to estimate where future performance-relevant errors are likely to occur, using this signal to guide exploration during MPC. In simulation, TERRA reaches approximately 90% slope-climbing success within five trials under severe out-of-distribution pretraining conditions, outperforming exploration baselines in reward, completion time, and energy to first success. On the ERNEST rover, TERRA recovers coordinated active-suspension locomotion on a deformable Mars-simulant slope within three to five real-world trials. Project website: https://terra-explore.github.io

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑