A CPG-based Control Architecture for 3D Locomotion of a snake-like robot A CPG-based Control Architecture for 3D Locomotion of a snake-like robot

In this paper, a biologically inspired control architecture for a snake-like robot is proposed to achieve 3D locomotion and realize continuously free gait transition. Based on a novel central pattern generator(CPG) model, which is achieved from the perspective of network synchronization, the control architecture integrates three functional parts. First, following the convergence behavior of the gradient system, a new CPG model is presented to adjust the signals’ amplitudes and phase differences automatically. Then, the relation characteristics between the CPG parameters and the outputs are investigated, which can be used to optimize the network’s efficiency and fault tolerance. Finally, to realize various stable gaits as well as free gait transition for snake locomotion, two kinds of gaits for snake-like robot are present with detailed phase difference parameters. The effectiveness of the proposed control architecture is demonstrated by 3D locomotion simulations and real-life experiments of a snake-like robot. The results show that desired locomotion patterns can be achieved after the CPG parameters are adjusted to corresponding gaits.