Central pattern generator based gait control for planar quadruped robots

In this paper, a gait control scheme is presented for planar quadruped robots based on a biologic concept, namely central pattern generator (CPG). A CPG is modeled as a group of the coupled nonlinear oscillators with an interaction weighting matrix which determines the gait patterns. The CPG model, mapping functions and a proportional-differential (PD) joint controller compose the basic gait generator. By using the duty factor of gait patterns as a tonic signal, the activity of the CPG model can be modulated, and as a result, a smooth transition between different gait patterns is achieved. Moreover, by tuning the parameters of the CPG model and mapping functions, the proposed basic gait generator can realize adaptive workspace trajectories for the robot to suit different terrains. Simulation results illustrate and validate the effectiveness of the proposed gait controllers.