Skip to content

论文arXiv 2609.342221 个月前

四足机器人本体感知力估计实现负载运动与牵引引导统一 ​

本文由 AI 根据公开内容整理,可能有误,请以原文为准:Proprioceptive Force Estimation for Quadruped Locomotion and Human-Robot Interaction(arXiv 2609.34222)

核心内容 ​

针对四足机器人负载适应与牵引交互需求,该研究提出共享三维力估计框架,利用本体感知历史推断持续载荷下的外力。通过联合训练力估计器与运动策略并学习潜在上下文,实现了单一力估计接口同时支持负载补偿运动与牵引速度/偏航指令生成。仿真测试显示分量力均方根误差 temporal mean 介于 1.44 至 2.83 N 之间。

相比域随机化基线,该框架使速度跟踪误差降低 21.6%,基底姿态误差减少 46.5%,持续受力测试平均存活率从 68.29% 提升至 94.60%。在 Unitree Go1 实机验证中,系统不仅实现静止态垂直/水平力估计,还成功完成超过 70 N 训练上限的 8.5 kg 负载运动,并复用同一力估计接口实现了稳定的牵引引导功能。

关键看点 ​

  • 联合训练架构:估计器与运动策略共享潜在上下文,以监督方式输出力与速度,解析映射生成平面速度与偏航率指令
  • 仿真性能提升:分量力RMSE为1.44-2.83N,较域随机化基线速度跟踪误差降21.6%,姿态误差减46.5%
  • 鲁棒性增强:持续受力仿真测试中平均存活率从68.29%显著提升至94.60%,验证策略对未知外力的适应性
  • 实机泛化验证:Unitree Go1实测支持8.5kg超载运动(超70N训练限),并复用同一力估计接口实现牵引引导

工程与产业观察 ​

该工作证明了仅依赖本体感知即可在无外部传感器条件下统一负载适应与人机牵引交互,大幅降低硬件成本。其联合训练与解析映射解耦思路,为具身智能在非标环境下的零样本力控迁移提供了高参考价值范式,尤其适用于无视觉或触觉传感器的低成本四足平台部署。

英文摘要 (Abstract) ​

Payload forces must be accommodated during locomotion, while leash forces can specify desired motion. We investigate whether a shared three-dimensional force estimate in newtons, inferred from proprioceptive history under sustained loading, can support both tasks. An estimator and locomotion policy are jointly trained with supervised force and velocity outputs and learned latent context. The estimated force conditions locomotion and additionally generates planar-velocity and yaw-rate commands for leash guidance through an analytical map. In sustained-force simulation sweeps, temporal means of componentwise force root mean square error range from 1.44 to 2.83,N. Compared with a domain-randomized baseline, the framework reduces velocity-tracking and base-orientation error scores by 21.6% and 46.5%, respectively, and increases mean survival from 68.29% to 94.60% in separate sustained-force tests. Unitree Go1 experiments demonstrate stationary vertical and horizontal force estimation, locomotion with an 8.5,kg payload whose weight exceeds the 70,N training force limit, and leash guidance using the same force-estimation interface.

作者:Run Wang, Xu Yang, Alapati Tuerxun, Yilin Mo

来源

本页事实对应的公开页面,以原文为准。

  1. arXiv 2609.34222:Proprioceptive Force Estimation for Quadruped Locomotion and Human-Robot Interaction访问 2026-10-12