Simulation Research On SLIP Based Trotting Motion Control Of Quadruped Robot

Currently, motion control is an important guarantee for high-speed locomotion andterrain adaptability of quadruped robot. High-speed dynamic locomotion of quadruped robotshows the characteristics of complexity, high-dimension, non-linearity, and dynamic coupling,bringing great challenges to the motion control of the robot. This thesis starts from a Bionicsview and designs the motion control and gait generation algorithms for quadruped robottrotting on both flat and uneven terrains based on Spring-Loaded Inverse Pendulum (SLIP),the equivalent model and control template of running.This thesis analyzes quadruped gait description methods and typical gait forms, focusesitself on the gait of running trot and analyzes its SLIP equivalent model, establishes thedynamic model of the SLIP system in flight and stance phase respectively with body mass andinertia considered, and deduces the stance-phase approximate solutions, providing thetheoretical base for understanding the dynamic features of SLIP model and for the design ofthe motion control algorithms.A SLIP-based motion control method is designed for quadruped robot trotting on bothflat and uneven terrains, including the swing angle control of the springy leg in flight phase,stiffness control of the springy leg in stance phase, and the adjustment of pitch attitude of thebody. With the concept of quadruped virtual springy leg and the equivalent condition of bothposition and force, the method is extended to the more complex and practical quadrupedsystem. The equivalent motion control and trotting gait generation algorithms are designed forquadruped robot through the planning of swing angle and retracting of the virtual leg in flightphase, and the torque control of the leg joints in stance phase.Adams-Simulink co-simulation platform is established to verify the effectiveness of thecontrol algorithms. Simulation results verify the effectiveness of the control algorithms forboth SLIP and multi-joint quadruped systems. The algorithms realize the effective control ofhorizontal speed, hopping height, and body pitch of the robot trotting on the flat, anddemonstrate good adaptability for uneven step terrains.