Study On Motion Control Of Underwater Robotic Fish Propelled By Pectoral/Caudal Fins

Underwater robot with mechanic propulsion is incommensurable with robotic fish which has high-efficient and high-mobility swimming method like the real fish.For this reason,underwater robotic fish as an important role in the field of robot is obtained the widespread attention from science and engineering.In this paper,the robotic fish that is propelled with 2-DOF pectoral fins and three-joint caudal fins is created and its motion control is the subject of this investigation.The main contents of this paper includes:(1)The sinusoidal motion law of each degree is given,and the hydrodynamic model based on turning characteristics of robotic fish is established.On this foundation,this article addresses turning characteristics on the surface and underwater of robotic fish.Turning on the surface is mainly divided into the fixed-radius turning with pectoral/caudal fins and spot turning with the couple of pectoral fins.The paper analyses the relationship between kinematics parameters of two different turning method(like period,amplitude,initial angle)and the turning radius,angular velocity using a numerical simulation method.Among them,maximum angular velocity of spot turning is 0.196rad/s and it is 0.78rad/s in fixed-radius turning.Turning underwater also called 3D turning includes fixed-depth turning and vertical turning in 3D space.Experiments show that robotic fish can accomplish the 3D turning motion.(2)The Hopfield neuron oscillator is used as each CPG(Central Pattern Generator)joint oscillator and then the two-ways coupling CPG model of seven joints of robotic fish is established.On this basis,the control conditions of the multiple modals which consists of swimming forward/backward,turning,pitching and rolling on the surface and underwater are obtained.The dynamic analyses of the pitching and rolling are studied in this paper.The reliability of CPG control and the maneuverability of robotic fish are theoretically and experimentally examined.(3)By using the vision sensor Pixy CMUcam5,a novel visual control system has been established to robotic fish visual control.The CMUcam5`s target tracking algorithm with color feature as target is described.Based on the above analysis of the surface and underwater motion of the robotic fish,the swimming modals of the robotic fish are divided and a tracking strategy is formulated.Finally,the fixed target and activity target tracking experiments are performed to verify the visual fish system based on CMUcam5.