NSGA-II Optimization-Based CPG Phase Transition Control Method of Manta Ray Robot

The manta ray robot switches the state of its pectoral fins to achieve different swimming states. In this paper, we propose a control method for pectoral fin flutter based on the central pattern generator (CPG). The pectoral fin state is determined by the amplitude, frequency, and phase difference of the CPG output signal. Momentary manipulation of the phase signal output from the CPG can produce a spike or interruption, which leads to unstable pectoral fin flutter of the manta ray robot. Too slow a phase transition can make the response slower and reduce the maneuverability of the robot. An optimization method for the phase transition of the manta ray robot’s pectoral fin state is investigated. With the output signal smoothness and rapidity as the optimization objectives, the Nondominated Sorting Genetic Algorithm-II (NSAG-II) is used to optimize the CPG parameters and obtain the optimal parameter combination. It is verified by simulation and experiment that the optimized CPG controller output signal can transition smoothly and quickly during phase switching.