在分析仿人机器人运动规划特性的基础上,对仿人机器人运动规划涉及的路径规划和步态规划两大问题及其典型方法进行了阐述和分析。对基于博弈论思想的离线足迹规划和基于传感信息融合的在线滚动路径规划两种路径规划策略进行了剖析,同时对几何约束法、模糊逻辑法、神经网络法、遗传算法法、自然步态法等5种常用的离线步态规划方法和3类在线姿态调整及控制方法即基于动力学模型的方法、基于倒立摆模型的方法、不基于模型的方法的算法思想和实验应用进行了分析与评价。最后对仿人机器人运动规划评价方法和运动规划研究的发展进行了讨论。
The paper surveys research advances in motion planning for humanoid robots, which has been a focus of robotics research in recent years. Features of motion planning for humanoid robots are discussed. Two aspects of humanoid motion planning, path planning and gait planning, are analyzed. Two path planning strategies, the game theory based footstep planning and the sensor based path replanning, are introduced. Five algorithms for offline gait planning, namely geometric constraint method, fuzzy logic method, neural network method, hereditary algorithm method and natural gait method, and three algrithoms for online gesture modification, namely dynamic model based method, inverted pendulum model based method and non-model based method, are introduced. Finally the assessment of motion planning for humanoid robots is discussed. Above analysis and estimation aim to provide a systematical overview on motion planning for humanoid robots.