Gait transition and locomotion control of a mantis-like robot based on improved neural coupled central pattern generator networks

Abstract Hexapod robots have been applied often due to their high flexibility, stability, and reliability. However, the control of hexapod robots is more complicated due to their highly redundant structures and intricate degrees of freedom. Inspired by the praying mantis, this paper presents the design of a hexapod mantis-like robot equipped with passive elastic legs. An improved neural coupling Central Pattern Generator (CPG) network, based on the Kimura oscillator, is developed for stride adjustment and locomotion control of the robot. A standardized function is designed to process the phase signals, which reduces the influence of the amplitude change of the phase signal in the adjustment area on the motion stability of the robot. The trajectory of the robot’s single-leg tip is planned, and a mapping function is designed to establish a relationship between the phase signals output by the CPG network and the single-leg tip trajectories, which enable the robot to adjust its stride and switch gaits by modifying the CPG network parameters. The feasibility of the proposed theory is verified by the prototype control experiments, which provide some inspiration for the control of other types of hexapod robots and have universal significance.