Myriapod robot locomotion control using central pattern generator based on Kuramoto oscillator model*

Online trajectory generation for myriapod robots with multiple degrees of freedom with neural control is still a difficult and unsolved problem. For this, we have designed a control architecture based on central pattern generators (CPGs) implemented as a system of coupled nonlinear oscillators of Kuramoto type. The CPG, like its biological counterpart, can produce coordinated patterns of rhythmic activity. To test our controller, we designed a simple myriapod robot with twelve actuated legs capable of walking on ground. A ring of 12 oscillators with projection of phase using a piecewise linear function serves as the CPG controller. Such a network running on specialized Tweensy 4.0 micro controller board takes around 24 kilo bytes of memory (only 0.2% of micro controller dynamic access memory capacity) to execute walking mode of locomotion. Results are presented demonstrating the basic locomotion pattern such as forward, right turning, and left turning movement of the robot and robustness of forward movement with respect to neural disruption of muted oscillators.