Locomotion generator for robotic fish using an evolutionary optimized central pattern generator

Abstract

Central Pattern Generator (CPG) consists of biological neural networks that generate coordinated rhythmic signals for the control of locomotion of vertebrate and invertebrate animals, such as walking, running, swimming and flying. In this paper, an evolutionary optimized CPG structure is proposed for generating fish-like locomotion of the robotic fish by controlling the flapping angles of all joints. The proposed CPG structure consists of three neural oscillators and each neural oscillator generates rhythmic signals for the corresponding joint of the three-joint robotic fish. The CPG structure for autonomous repeated locomotion has the parameters which determine the form of output signals. Quantum-inspired Evolutionary Algorithm (QEA) is employed for optimizing these parameters to generate signals which track the kinematically derived fish-like locomotion. The effectiveness of the proposed CPG structure is demonstrated by computer simulations.