CPG-based gait transition of a quadruped robot

Quadruped animals have high agility and grace skills in the motion over natural environment. An important feature of such animals is their abilities in the selection of appropriate gait in the response to the terrain irregularities and also the smooth transition between them in order to increase the stability or conserve the energy. In this paper, a Cartesian CPG-based controller for the gait generation and the transition between them will be developed. The proposed approach uses nonlinear coupled oscillators for generating rhythmic signals. These patterns can be modulated by a programmed excitation signal and proportionally changing the frequency, amplitude and the coupling parameters among oscillators. According to these variations, the step length, duration of swing and stance phases of all legs, relative phase between the legs and the sequence of leg lifting can be adjusted smoothly especially by the duration of gait transition. Then, trajectories of swing legs along the z- and x-axeis will be computed through the data obtained form the CPG controller. The reference ZMP path will be desgined on the basis of the footprints of legs. Then, the COG trajectories will be calculated with a preview servo controller. Finally, some simulations are conducted to verify the effectiveness of the proposed approach both in the gait generation and smooth gait transition for a quadruped robot.