Research On Closed-loop Motion Control Of Bionic Robot Fish With 3-DOF Pectoral Fin Propulsion

As a new type of underwater robotic that combines with fish propulsion mode and robotic technology,bionic robotic fish has attracted many researchers into this field because of its characteristics of high propulsion efficiency,strong maneuverability,and low impact on the environment.In order to make the bionic robotic fish have better water movement ability,its motion control problem has always been the focus of the research of the bionic robotic fish.Based on the bionic robotic fish platform independently developed by our laboratory,which has 3-DOF pectoral fin mechanism,this paper studies the motion control of bionic robotic fish.The main research contents are as follows:(1)To solve the question of gait generation and conversion of bionic robotic fish,we use a central pattern generator(CPG)to control each joint of robotic fish.First,a set of linear state equations and a sine output equation are used to construct a CPG oscillation unit,and the phase difference parameters are used to complete the coupling between the oscillators to form a CPG network.Then through the dynamic analysis and modeling of the robotic fish,the gait parameters corresponding to the different swimming modes are determined.The simulation results show the effectiveness of the network in gait control of the robotic fish and the flexibility in gait transition.(2)To solve the question of the pitch attitude control of the bionic robotic fish,we design a control method with the established CPG network and dynamic model.First,Taylor's formula is used to linearize the dynamic model to obtain a simplified pitch dynamic equation,and the control law framework is designed based on this.Secondly,for the problem of unknown control law parameters,the RBF neural network is used to estimate the unknown parameters to replace the unknown parameters.The stability of the control law is proved by Lyapunov's second theorem,and the RBF weights are derived.Adaptive law.Finally,according to the control law,the relevant parameters of the pectoral fin CPG are adjusted to achieve the pitching motion.Simulation results show that under this control law,the robotic fish can adjust the pitch angle to the desired position.(3)To solve the problem of obstacle avoidance motion control of the bionic robotic fish,we propose a collision-cone-based method.The obstacle avoidance behavior of robotic fish is divided into two cases: deflecting obstacle avoidance and pitching obstacle avoidance.First,according to the collision cone theory,get the condition of colliding with obstacles of the robotic fish,and determine which way to avoid obstacles.Then,according to the given control method,get the pitch(deflection)angle of the robotic fish.Finally,according to the deflection angle and the result of the control law,the CPG network control parameters are changed to achieve obstacle avoidance behavior.Simulation results show the effectiveness of the control strategy.