On Study of Intelligent Control Methods for a Bionic Dolphin Based on Sensory Feedback CPG Model

摘要

Widely applied in motion controllers, the central pattern generator (CPG) model is able to generate multiple swimming patterns by the same CPG network, but is poor in stability and adaptability. Therefore, the studies on methods of closed-loop sensory feedback CPG are insufficient. While in this paper, the author proposes a new artificial intelligence control model for a bionic dolphin based on the closed-loop sensory feedback CPG, which is suitable for self-propelled swimming underwater of multi-joint bionic dolphin. The model is composed of analogous lamprey oscillator and a sensory neuron (SN) one-way link network, and it provides oscillation output with stable rhythm for multi-joint bionic dolphin, and is equipped with a sensory feedback mechanism. Moreover,a decision-making mechanism-fuzzy logic control is designed to make it perform through a finite state machine (FSM) algorithm when autonomously executing a given task. According to the experimental analysis on still waters, the model shows good self-propelled swimming performance and high yaw control accuracy in obstacle areas.