CPG-based autonomous swimming control for multi-tasks of a biomimetic robotic fish

Abstract

This work emphasizes the artificial intelligent control of the developed biomimetic robotic fish (i-RoF) based on sensory feedback Central Pattern Generator (CPG) approach. In order to obtain rhythmic, robust and adaptable closed loop control performance, a novel control mechanism, composed of two subsystems; as a biological based CPG network and a Fuzzy Logic controller, is proposed. The CPG network behaves as an artificial spinal cord, inspired by Lampreys. It is constructed as unidirectional chained network with Sensory Neurons (SNs) which can perceive the external stimulus. The Fuzzy Logic control structure is also designed as a decision-making mechanism and a Finite State Machine (FSM) algorithm is proposed to perform the given missions autonomously. In the experimental studies, different scenarios are created for both situations as autonomous swimming and encountering obstacles while taking on a primary task, which is determined with yaw control. Autonomous swimming ability is performed in the real pool environment with proposed scenarios that are likely to be encountered in a real environment. It is observed that autonomous swimming performance is excellent and yaw control is quite good even in the presence of obstacles. It is also proved that the escape performance of the robot is very fast.