Gait Generation With Smooth Transition Using CPG-Based Locomotion Control for Hexapod Walking Robot

摘要

This paper presents a locomotion control method based on central pattern generator (CPG) for hexapod walking robot to achieve gait generation with smooth transition. By deriving an analytical limit cycle approximation of the Van der Pol oscillator, a simple diffusive coupling scheme is proposed to construct a ring-shape CPG network with phase-locked behavior. The stability of the proposed network is proved using synchronization analysis with guaranteed uniform ultimate boundedness of the synchronous errors. In contrast to conventional numerical methods in tuning parameters of the CPG network, our method provides an explicit result that could analytically determine the network parameters to yield the rhythmic waveforms with prescribed frequency, amplitude, and phase relationship among neurons. Employing the proposed network to govern the swing/stance phase according to the profile of the resulting CPG signals, a locomotion control strategy for the hexapod robot is further developed to manipulate the leg movements during the gait cycle. By simply adjusting the phase lags of the CPG network, the proposed control strategy is capable of generating stable walking gaits for hexapod robots and achieving smooth transition among the generated gaits. The simulation and experimental results have demonstrated the effectiveness of the proposed locomotion control method.