欠驱动单腿机器人膝踝协调运动研究

发布时间：2018-04-20 15:41

本文选题：单腿机器人 + 膝踝协调运动　；参考：《浙江大学》2017年硕士论文

【摘要】：腿足机器人具有良好的运动性能和地形适应能力,一直是机器人领域研究的热点。单腿机器人是腿足机器人的一个重要分支,研究其跳跃机理可以为仿人机器人和多足机器人的快速跑跳运动提供理论依据。目前国内外单腿机器人仍以两自由度关节为主,机器人的运动性能和能量效率受到限制。本文基于仿生思想,在单腿机器人中加入脚踝关节,采用SEA(Series Elastic Actuator)作为关节执行器。这样不仅可以提高机器人跳跃运动的能力,还可以减缓机器人落地时的冲击力并存储跳跃过程中的能量,提高机器人跳跃运动的稳定性和能量效率。但是,随着机器人关节数量的增加,机器人的运动规划和控制难度也将增加。本文面向基于SEA关节的三段式单腿机器人,以提高机器人的跳跃性能和能量效率为目标开展研究,研究内容分为以下方面:1.运动规划与控制算法。本文针对机器人的多关节冗余问题,在飞行相中,设计逆运动学优化算法;在站立相中,结合人体的运动学和动力学规律,以能耗最小为目标,定义膝踝协调的评价指标,设计运动学规划算法,并设计弹簧控制率从而控制机器人站立相的膝踝关节运动。2.控制参数优化算法。本文针对弹簧控制率中膝踝关节的弹簧刚度选取问题,以膝踝协调评价指标最高为优化目标,采用PSO算法优化膝踝关节的弹簧刚度;提出变刚度的能量补偿算法,补偿机器人落地冲击时的能量损耗。3.机器人跳跃运动控制实验。本文分别在机器人仿真平台和实物机器人上开展了实验。在仿真平台中,通过PSO算法的刚度优化结果和变刚度能量补偿算法,对机器人前向跳跃运动进行了研究。在被动脚踝机器人中,本文根据PSO算法与优化结果,选取了机器人的踝关节弹簧。并采用变刚度的能量补偿算法,对机器人的连续跳跃运动进行了控制,实现的最大跳跃高度为35cm,最高单位距离能耗为0.29。在主动脚踝机器人中,优化了踝关节的弹簧刚度,对单向SEA驱动的脚踝关节进行了位置和力矩控制,开展了连续跳跃实验。仿真和实验结果均验证了本文所提规划和控制方法的有效性。
[Abstract]:Legged robot with good performance of motion and terrain adaptation has been a hot spot in the field of robot research. One-legged robot is an important branch of leg-legged robot. The study of its jumping mechanism can provide a theoretical basis for the rapid movement of humanoid robot and multi-legged robot. At present, the single leg robot is still mainly two degree of freedom joints at home and abroad, and its motion performance and energy efficiency are limited. Based on the bionic theory, the ankle joint is added to the single legged robot, and SEA(Series Elastic Actuator is used as the joint actuator. This can not only improve the jumping ability of the robot, but also slow down the impact force and store the energy in the jumping process, so as to improve the stability and energy efficiency of the robot jumping motion. However, with the increase of the number of robot joints, robot motion planning and control will be more difficult. This paper aims at improving the hopping performance and energy efficiency of the three-segment single-legged robot based on SEA joint. The research contents are as follows: 1. Motion planning and control algorithm. In this paper, the inverse kinematics optimization algorithm is designed in flight phase to solve the problem of multi-joint redundancy of the robot. In standing phase, with the aim of minimum energy consumption, the evaluation index of knee and ankle coordination is defined in the standing phase, combined with the kinematics and dynamics of human body. The kinematics planning algorithm is designed and the spring control rate is designed to control the motion of the knee and ankle joint of the robot standing phase. Control parameter optimization algorithm. In this paper, aiming at the problem of spring stiffness selection of knee ankle joint in spring control rate, PSO algorithm is adopted to optimize spring stiffness of knee ankle joint with the highest evaluation index of knee and ankle coordination, and an energy compensation algorithm with variable stiffness is proposed. Compensates the robot ground impact energy loss. 3. Robot jumping motion control experiment. Experiments are carried out on the robot simulation platform and the real robot. Based on the PSO algorithm and the variable stiffness energy compensation algorithm, the forward jump motion of the robot is studied in the simulation platform. In the passive ankle robot, according to the PSO algorithm and the optimization results, the ankle spring of the robot is selected. The continuous jump motion of the robot is controlled by the variable stiffness energy compensation algorithm. The maximum jump height is 35 cm and the maximum energy consumption per unit distance is 0.29. In the active ankle robot, the spring stiffness of the ankle joint is optimized, the position and torque control of the ankle joint driven by unidirectional SEA is carried out, and the continuous jumping experiment is carried out. Simulation and experimental results verify the effectiveness of the proposed planning and control methods.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】