Moving control of quadruped hopping robot using adaptive CPG networks

This paper describes the moving control using the adaptive Central Pattern Generators (CPGs) including motor dynamic models for our developed quadruped hopping robot. The CPGs of each leg is interconnected with each other and by setting their coupling parameters can act as the flexible oscillators of each leg and adjust the hopping height of each leg to require stable hopping motion. The formation of the CPG networks are suitable not only to generate the continuous jumping motion but also can generate the moving motion in two-dimensional, respectively. We also propose the reference height control system which including the maximum hopping height detector and Proportional Integral (PI) controller to achieve the reference jumping height. By using the proposed method, the hopping height of each leg can be control independently in order to make the posture of robot's body incline ahead and move forward. We create MATLAB/Simulink model to conduct various types of experiments and confirmed the effectiveness of our proposed CPG model including the reference height control system to generate the stable moving performance while jumping continuously.