Gait planning of a quadruped robot integrating inverse kinematics and CPG
Long Huang,Xuan Wu,Xiaojie Wang
DOI: https://doi.org/10.1117/12.2582535
2021-01-01
Abstract:In this paper, we designed a quadruped, dog-like robot which has 3 degrees of freedom (DOFs) per leg, and combined two commonly used methods for robot’s gait control, to make better gait planning strategy in different environments. First, we chose the inverse kinematics control method, and established the quadruped robot simulation motion model in Webots to verify the validity of the motion modes such as forward, backward, turning, and height adjustment of the body. Then we utilized the CPG (Central Pattern Generator) control method, and an improved Hopf nonlinear harmonic oscillator, which requires less parameters, smaller computation complexity, and clearer physical meaning, to generate a faster trot gait. According to the timing diagram of different gaits, the coupling topology structure between CPG units was designed, with joint motion curve generated. The simulation and experiment show that compared to inverse kinematics, CPG is more suitable for producing simpler and faster periodic gait planning signals, while inverse kinematics method takes advantages in the body’s self-balance and foot’s adaptation on the ground. The two methods can be complementary to the motion control of the quadruped robot.