外观
论文arXiv 2610.2024114 天前
利用几何候选模式从多模态示范中学习导航
原文链接：Learning Navigation from Multimodal Demonstrations with Geometric Candidate Modes（arXiv 2610.20241 · 发布日期：2026-10-14）
Learning Navigation from Multimodal Demonstrations with Geometric Candidate Modes
针对离线导航示范多模态导致行为克隆取平均的问题，提出基于局部几何候选模式的分层策略，并在仿真和LimX Tron 1双足机器人上验证。
核心要点
- 离线导航中同一观测可能对应多种有效动作，行为克隆会平均这些动作并可能产生不安全路线。
- 方法从局部几何计算候选模式，高层策略结合目标和上次选择来选模式，低层策略将其转为动作。
- 仿真中在已见和未见测试案例成功率超过95%，高于四个离线学习基线和同模式启发式方法。
- 仿真训练的策略还迁移到LimX Tron 1双足机器人，摘要未提到开源信息。
验证范围与局限：验证范围包括仿真中的已见和未见测试案例，以及一台LimX Tron 1双足机器人。
英文摘要 (Abstract)
Offline navigation demonstrations are multimodal, because one observation often allows several valid actions, such as passing an obstacle on either side. Behavior cloning averages these actions, potentially leading to unsafe routes. Multimodal policies must discover the alternatives from the data, but similar states can lead to different routes whose current actions overlap. In navigation, the alternatives have a geometric cause and can be computed from one observation. We propose a hierarchical neural-network policy that computes candidate modes from local geometry and learns from demonstrations which mode to follow and how. A high-level policy selects a mode using the goal and its previous choice among a set of candidate modes, and a low-level policy turns it into an action. In simulation, the policy succeeds in over 95% of seen and unseen test cases, compared with at most 43% and 20% for four offline learning baselines and 79% and 67% for a heuristic that uses the same modes without learning. The simulation-trained policy also transfers to a LimX Tron 1 bipedal robot.
作者：Chengyang Peng, Zhihao Zhang, Upama Pant, Yiming Wang, Keith Redmill, Ayonga Hereid
来源
本页事实对应的公开页面，以原文为准。