Closed-Loop Pitch Attitude Control of Biomimetic Robotic Fish

Abstract

This paper considers the pitch control of a robotic fish with three degree-of-freedom(DOF) pectoral fins, wire-driven flexible body and a passive caudal fin. The linear model with undetermined parameters is first suggested to describe the dynamics of the pitch angles, which are the control parameters to adjust the posture of robotic fish. Then, a control network based on Central Pattern Generators(CPGs) is proposed to separately adjust the state of each DOF. Finally, a compound controller is designed to adjust the amplitudes of pectoral Fins such that the desired pitch angle is tracked, in which an adaptive law combined with RBF neural network is proposed to estimate the undetermined parameters. The stability of the proposed algorithm is proved and then the validity is investigated by the simulation results.