Parameter Region Analysis And Optimization For Cycle Gait In Compass-like Model

The study of humanoid robot mainly focuses on human-like walking,and the stability of the robot motion and energy dissipation problem are very important for its intelligent development.Currently,the way of active dynamic walking is often applied,which is of high control precision,but costs too much,and with heavy weight.The quasi-passive dynamic walking with swing and touch the ground switching process,without control among swing,is natural and in lower energy consumption.However at present,the passive dynamic walking is mainly used for the analysis of the complex gait and the theoretical control algorithm,but with very little consideration of the influence of parameters on the system.The Compass-like model,proposed by Goswami,is the most representative model of the human leg structure in the passive walking model.In this paper,the study are given on the parameter domain and the domain of attraction from the combination of theoretical analysis and numerical calculation based on the previous studies on the Compass-like model.1)The theoretical analysis of the Compass-like model system is made,including the dimensionless parameters and the system parameters.The authors analyze the stability of the passive walking motion by using the idea of Poincare map,and study the relationship between the stability of a single parameter and the stability of the model.2)This paper describes the way of finding the stable parameter region of the model,and uses the method to find the parameter region of 2D and 3D.Extraction of domain parameter generation into the system equation,and verifies the feasibility and reliability of the search algorithm by analyzing the system phase diagram.This paper also analyzes the influence of different values in different parameters on the domain of attraction.3)According to the changing trajectory near the switch surface of the Jacobian matrix,this paper achieves the compensation system of Lyapunov Exponent,judging whether the Compass-like model appears chaotic.Using "dimensionality reduction-rising dimension",and the Matlab toolbox,it introduces the topological search in detail,and find a threedimensional topological horseshoe periodic chaotic gait in phase space,which determines the existence of chaotic gait.In this paper,we simplified the Compass-like passive walking model and analyzed the influence of each parameter on the stability of the model.In this case,we optimized the attractor region and proposed a method of researching global stability parameters.And we used the Lyapunov Exponent method and topological horseshoe theory to demonstrate our research results.