Coordinated Motion Control Between The Knee And Ankle Joints For Under-actuated One-legged Robot

Legged robots have been a popular research focus in robotics field,owning a favorable motion performance and terrain adaptability.As one of the most important branches of legged robots,the research on the jumping mechanism of one-legged robots provide a strong theoretical support for humanoid robots and multi-legged robots for their fast running and jumping.Nowadays,the research on one-legged robots is still mainly focused on the robots with two-DOF joints,which heavily limits the motion performance and energy efficiency.In this thesis,the ankle joint is added to the robots based on the bionic idea,and the SEA(Series Elastic Actuator)is used as these joint actuators.It will not only improve the jumping motion ability but also reduce the impact force and help store energy during landing.Furthermore it increases the stability of robots'running and jumping motion as well as energy efficiency.However,as the number of joints increases,the motion control and planning will become more difficult.This thesis focuses on the three-segment one-legged robot with SEA joints.In order to improve jumping motion performance and energy efficiency,the research contents are as following.1.Motion planning and control algorithm.In order to solve the multi-joints redundant problem of the robots,an inverse kinematics optimization algorithm is introduced in flight phase.In standing phase,in order to minimize energy consumption,we define the coordination evaluation index between knee and ankle joints combining with the human kinematics and dynamics principle.The kinematic planning is designed and a Spring-Control Method is proposed to control the motion coordination between knee and ankle joints.2.Control parameter optimization algorithm.For the purpose to select the spring stiffness for knee and ankle joints according to the Spring-Control Method and to maximize the coordination evaluation index between them,this thesis adopts PSO(Particle Swarm Optimization)algorithm to optimize the stiffness of knee and ankle joints.Besides,an energy compensation algorithm with variable stiffness is proposed to compensate the energy loss during landing.3.Experiments in jumping motion control.Experiments are conducted both in simulation and physical robots.In simulation,the research is focused on forward jumping using the stiffness optimization results of PSO algorithm and variable stiffness method.For the robot with passive ankles,an ankle joint spring based on PSO algorithm and the optimization results is elected.Continuous jumping motion is successfully achieved with the maximum jumping height which is 35cm and the maximum energy consumption per unit distance is 0.29 under variable stiffness method.For the robot with active ankles,the spring stiffness of the ankle joint is optimized.The ankle joint of one-way SEA is fully controlled on position and torque.Continuous jumping motion experiments are also conducted.The experiment results prove the effectiveness of the algorithm proposed in this thesis.