外观
论文arXiv 2610.1961513 天前
PolyMorph框架实现单一策略跨双足三足四足机器人通用运动控制
本文由 AI 根据公开内容整理，可能有误，请以原文为准：PolyMorph: Learning Locomotion Policy for Cross-Embodiment Adaptation（arXiv 2610.19615）
核心内容
针对腿足机器人运动策略难以跨形态复用的痛点，研究提出PolyMorph框架，旨在训练单一通用运动策略以适配异构机器人。该方法将URDF解析为关节-连杆模块，结合循环编码器、图Transformer及逐关节解码器生成动作。实验在仿真中覆盖双足、三足、四足及未见变体，并成功部署于真实四足平台多种构型，证实单策略可泛化至多样形态且无需重训。
技术层面，PolyMorph采用形态归一化奖励函数、自适应机器人采样及基于URDF的PD增益初始化以提升训练稳定性。系统架构核心在于利用图结构处理可变腿数与关节拓扑，通过形态条件输入动态调整策略输出。该方案解决了传统RL策略对特定执行器数量与机械属性的强耦合问题，为具身智能跨本体迁移提供了标准化的算法范式与工程验证路径。
关键看点
- 图Transformer架构：结合循环编码器与逐关节解码器，将URDF转为关节-连杆模块，支持可变腿数与关节配置的动态动作生成
- 训练稳定性优化：引入形态归一化奖励、自适应机器人采样策略及URDF衍生PD增益初始化，解决异构机体联合训练收敛难题
- 跨形态泛化验证：单一策略在仿真中覆盖双足、三足、四足及未见构型，并在真实四足平台多配置下完成零样本部署
- 解耦形态依赖：突破传统RL策略对特定执行器数量与机械属性的绑定，实现无需针对新机体重新训练的通用运动控制
工程与产业观察
PolyMorph验证了基于图结构与形态条件化的RL框架可实现跨本体运动策略复用，大幅降低异构机器人适配成本。其URDF模块化解析与归一化训练机制为具身大模型提供底层运动基座参考，推动从“一机一策”向通用运动原语演进，加速多形态机器人研发迭代效率。
英文摘要 (Abstract)
Reinforcement learning has enabled robust locomotion for legged robots, but learned policies are typically tied to a specific robot morphology and require additional engineering when the number of legs, actuators, or mechanical properties change. We present PolyMorph, a morphology-conditioned framework that learns a single locomotion policy for heterogeneous legged robots. PolyMorph converts each robot's URDF into joint-link modules and combines a recurrent encoder, graph-based Transformer, and per-joint decoder to generate actions across variable leg counts and joint configurations. To improve training stability, PolyMorph uses morphology-normalized rewards, adaptive robot sampling, and URDF-derived PD-gain initialization. We evaluate PolyMorph in simulation across bipedal, tripedal, and quadrupedal robots, including unseen morphology variants and platforms, and deploy the policy on real quadruped platforms operated in multiple configurations. The results show that a single PolyMorph policy generalizes across diverse robot morphologies without per-robot policy retraining.
作者：Hyunsik Oh, Seunghun Jeon, Donghoon Youm, Jungwoo Hur, Hyeongjun Kim, Jeong Hyun Lee, Moonkyu Jung, Jemin Hwangbo
来源
本页事实对应的公开页面，以原文为准。