Model Analysis And Motion Control Research Of Biped Robot

The biped movement is characterized by complex system model and difficulty in engineering implementation.This field covers the challenging problems including hybrid nonlinear models,underactuated systems,environmental interactions and other related disciplines.It is a representative problem in the field of robots from the perspective of theory and application.Dimension disaster caused by the discretization of biped robot's continuous systems makes it difficult to find the complete solution to the problem.Therefore,researchers and scholars have been dedicated to find effective methods.In recent years,with the continuous advancement of engineering and technologies,some developments in theoretical algorithms have given rise to numerous excellent outcomes,bringing research in this field to a new hot spot.Based on the current development in biped robotics field,this study includes four problems: simplified model of biped robot,whole body motion model of biped robot,joint actuation control and biped robot prototype development.The low-dimensional simplified model of the biped robot can reflect the essential characteristics of the biped robot movement to facilitate the motion planning in high level.This study analyzes three representatives of point mass models.Firstly,the selfstability of the inverted pendulum model is analyzed and the self-stability of the passive model with different mechanism configurations is studied.The relationship between the self-stability of the mechanical configurations of several legged robots and the inverted pendulum model is compared.Secondly,the linear inverted pendulum model is analyzed and the model predictive control technique based on linear inverted pendulum model is studied.Within the model prediction control framework,the control method of adjusting the yaw angular momentum by the swing leg inertia is proposed.Finally,the hypothesis that generalized point mass model is piecewise linearized is released.A sliding mode controller is designed and the robustness map is presented in the phase space,which extends the robustness of the existing prismatic inverted pendulum model.The high-dimensional full-body motion model of the biped robot maps the motion planning trajectory to the joint space,giving the joint moment or position needed to achieve the planned motion in high level.In this study,by adding the virtual degree of freedom in the floating coordinate on the biped robot,the underactuated system is transformed into the fully-actuated system.This study deduces the analytical solution to inverse dynamics based on the operational space framework.The advantage of this framework is that the inverse dynamics problem can be solved in real time,the task priority is clear,and it is convenient to integrate with the simplified point mass model in high level.In this study,the real-time simulation based on the operational space framework is carried out to study the stable motion control of a biped robot model with passive ankle joints.The effectiveness of the framework is verified.The biped robot whole body model solves the joint torques.To tackle the footground collision of biped robot,this study develops hydraulic SEA to realize variable stiffness control,designs a planar biped robot experiment platform based on hydraulic SEA,and embodies the spring-loaded inverted pendulum model.Two variable stiffness control methods are proposed for the joint nonlinear mechanism of the developed platform to verify the effectiveness of hydraulic SEA variable stiffness control for the foot-ground collision problem.A hydraulic-actuated biped robot prototype is developed.The design of the robot joints' transmission mechanism is optimized and key mechanical parts are analyzed.Hydraulic actuator has high output force,high stiffness and outputs linear motion.To meet the requirement of real-time control and integration of electrical systems,multithreads architecture is designed.In view of the high hydraulic pressure of the hydraulic system,the mechanical design of the hydraulic cylinder and the wiring mode of the hydraulic oil pipe are optimized.The experimental research on model predictive control based on linear inverted pendulum within angular momentum control embedded is carried out to verify the effectiveness and deficiency of the biped robot prototype platform performance and motion control algorithm.