仿人机器人刚-柔-软多体耦合动力学建模与全身协调控制方法研究

This project is going to perform the study of dynamic modeling and whole-body control of humanoid robot with rigid-flexible-soft structure. Based on the bionics of human torso, a flexible torso of humanoid robot consisting of rigid-flexible-soft structure will be designed and its performance analysis and structural optimization will then be carried out. By developing the modeling method of the flexible and soft structure in robot system, a dynamic model of the whole robot system with rigid-flexible-soft structure will be established and used for dynamic analysis. By establishing the dynamic model of the humanoid biped and the whole-body control algorithm under multi-contact condition, the humanoid robot will be able to complete tasks while maintaining its own stability. This project aims at the forefront of international research field of humanoid robot, with a focus on the cross of multi-disciplines, and hopes to make breakthroughs of rigid-flexible-soft structure design and dynamics, and the whole-body control of humanoid robot by interaction with environment. Rich fruits of this project will provide technical support and guarantee for the integration of humanoid robots into human life and working environment, as well as promote the combination of basic research and its application.

本项目拟开展仿人机器人刚-柔-软多体耦合动力学建模与全身协调控制方法研究。基于人体躯干仿生学，设计基于刚-柔-软耦合结构的仿人机器人柔性躯干，并进行性能分析与结构优化；发展机器人系统柔性与软体结构的建模方法，建立仿人机器人刚-柔-软多体耦合系统的动力学模型，并开展其动力学分析；建立仿人机器人的多点接触动力学模型，利用环境接触发展全身协调控制算法，实现仿人机器人任务操作与稳定性能的协调控制。本项目瞄准国际仿人机器人的研究前沿，注重多学科交叉，重点突破仿人机器人刚-柔-软耦合柔顺结构设计与动力学分析方法，以及仿人机器人利用环境接触的全身协调控制，为实现仿人机器人的人-机-环境共融提供技术支撑和保障，促进基础研究和应用研究的结合。

本项目瞄准国际仿人机器人的研究前沿，意在突破仿人机器人刚-柔-软耦合柔顺结构设计与动力学分析方法，以及仿人机器人利用环境接触的全身协调控制，为实现仿人机器人的人-机-环境共融提供技术支撑和保障。本项目在计划期内开展仿人机器人刚-柔-软多体耦合动力学建模与全身协调控制方法研究，进行刚-柔-软耦合仿人机器人结构设计，建立刚-柔-软多体耦合系统的动力学模型并进行仿真，据此提出了与仿人机器人周期步态规划、步态切换策略及不确定扰动下的鲁棒性有关的多种理论与控制方法。以此为基础，通过搭建机器人样机，建立多传感器融合采集系统，感知周围动态环境，进行了一系列仿人机器人路径规划与避障测试、在非平整路面等多种环境下的共融实验。已发表期刊论文8篇，其中有5篇已为SCI收录（1篇为网络出版,https://doi.org/10.1002/ asjc.2437），3篇为EI收录。已申请国家发明专利17项（其中已获授权10项），并登记2项软件著作权。