Rough Terrain Of Quadruped Robot Based On Nonlinear Optimization Research On Pose Optimization

Compared to other types of land robots,quadruped robots have the advantage of being able to traverse rough terrain with varying slopes or heights.They have been widely used in legged robots due to their structural and economic advantages.However,due to the uncertainty and complexity of the terrain environment,the stable traversal of quadruped robots over rough terrain is still a challenging research topic.Therefore,the research on pose optimization of quadruped robots has important value.This paper aims to improve the motion stability of quadruped robots in rough terrain,and explores the problem of optimizing the trunk pose of quadruped robots in rough terrain.The main work includes four aspects: kinematic modeling,crawling gait planning,center of gravity(COG)pose optimization,and simulation analysis.Specific research content and results include:1.Quadruped robot pose description and modeling.The characteristics of rotation matrices,Euler angles,and unit quaternions in representing attitudes are compared,and the attitude description method and transformation relationship of unit quaternions are introduced.The kinematic model of the robot is established,and its kinematic equations are obtained using the D-H method.2.Crawling gait planning.The methods for measuring the stability of quadruped robots are discussed,including the COG-based static stability criterion and the ZMP-based stability criterion.The calculation method of stability margin is further determined based on stability and walking efficiency.Based on stability and walking efficiency,the leg lifting sequence of the robot is planned,and the stability of the swing phase switching process is considered by adding the quadruped support state.The Hopf harmonic oscillator is used to simulate the CPG periodic oscillation signal to drive the leg movement of the quadruped robot,and the specified gait is generated.Finally,the swing phase end trajectory adapted to rough terrain is generated using Bézier curves.3.Quadruped robot pose optimization.A strategy for adjusting the COG position and posture of quadruped robots in rough terrain is proposed.The physical problem is converted into a mathematical problem using a nonlinear optimization problem(NLP).The pose adjustment strategy is mapped to the objective function,while considering stability and single-step maximum walking distance as constraints.The sequential quadratic programming(SQP)algorithm is used to improve the speed of solving nonlinear optimization problems.According to the relationship between the torso and feet in the continuous state of the robot,a nonlinear least squares problem is constructed to solve the initial value of the optimization problem.Hermite curves are used to fit the planned pose points to generate the motion trajectory of the robot trunk pose,and the default foot trajectory is mapped to the optimized trunk state.Finally,feedback control is used to complete the planned trajectory motion.4.The simulation experiment was designed with varying height and slope terrain using Adams and MATLAB joint simulation.The simulation results show that the quadruped robot can traverse steps,slopes,and stairs stably according to the planned trajectory,which verifies the reliability and effectiveness of the pose optimization method.