A backward control based on sigma-Hopf oscillator with decoupled parameters for smooth locomotion of bio-inspired legged robot

摘要

In order to move, animals in nature usually use a rhythmic movement which is highly stable and adaptable. Movement controlled by the central pattern generator (CPG) is a spontaneous behavior of the lower nervous center. When applying the CPG model to a multi-legged robot, the stability of the robot motion and the generation and transformation of the adaptive motion model are affected by the output of the CPG. Considering the challenges of strong coupling, complex controlling, and the difficulty of using the traditional CPG model to control the foot trajectory of the robot, a novel sigma-Hopf harmonic oscillator with decoupled parameters is proposed. In order to increase the flexibility and reduce the complexity of controlling the robot, a Central Pattern based Backward Control (CPBC) method is proposed. This ensures that the gait, duty factor and frequency control of the robot can be individually controlled. By adjusting the corresponding parameters, the robot can freely transform between the different motion states. Simulation techniques, where the parameters are dynamically adjusted, are used to test the effects of this process on the robot. A bio-inspired legged robotic platform is designed and tested to verify the CPBC method. (C) 2018 Elsevier B.V. All rights reserved.