arXiv cs.AI· Oliver Hausd\"orfer, Alexander von Rohr, \'Eric Lefort, Angela Schoellig·· 12 小时前AI 评分31
利用协同动作与风格先验学习足式机器人运动
Synergistic Action and Style Priors for Learning Legged Locomotion
AI 导读
研究者提出一种面向力矩控制的归纳偏置：从少量专家演示数据中学习潜在动作先验，使策略能直接利用专家动作知识并提升探索效率。该智能体不受演示奖励水平限制，迁移任务性能显著提升；潜在动作先验结合模仿风格奖励可更接近地复现专家行为。
正文 · 原文
Abstract:Deep Reinforcement Learning (DRL) enables robots to learn complex behaviors through interaction with the environment. However, due to the unrestricted nature of the learning algorithms, the resulting solutions are often brittle and appear unnatural. This is especially true for learning direct joint-level torque control, as inductive biases are difficult to integrate into the learning process. We propose an inductive bias for learning locomotion that is especially useful for torque control: latent actions learned from a small dataset of expert demonstrations. This prior allows the policy to directly leverage knowledge contained in the expert's actions and facilitates more efficient exploration. We observe that the agent is not restricted to the reward levels of the demonstration, and performance in transfer tasks is improved significantly. Latent action priors combined with style rewards for imitation lead to a closer replication of the expert's behavior. Videos are available at this https URL.
| Comments: | Humanoids 2026 |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2410.03246 [cs.RO] |
| (or arXiv:2410.03246v3 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2410.03246 arXiv-issued DOI via DataCite |
Submission history
From: Oliver Hausdörfer [view email]
[v1]
Fri, 4 Oct 2024 09:10:56 UTC (6,083 KB)
[v2]
Sat, 1 Mar 2025 09:12:55 UTC (8,986 KB)
[v3]
Tue, 20 Oct 2026 06:52:02 UTC (2,796 KB)
来源：arXiv cs.AI · arxiv.org