Structure Design Of Underwater Robot Manipulator And Research On Adaptive Neural Network Control

With the development of human society,the non-renewable resources on land have been gradually reduced.The ocean,as a treasure trove of abundant resources,has become the object of competitive research and development around the world.The development of the ocean requires advanced technology and equipment,and underwater robots have received much attention as important equipment for marine development.With the acceleration of the ocean development process,new requirements have been put forward on the function of the unmanned cable-free autonomous underwater vehicle(AUV),which is required to have a working function based on the detection function.Therefore,it is necessary to mount a robot on the AUV.The working type AUV is constructed,and the robot and the AUV carrier are cooperatively controlled to complete the work task.Due to the robot,the posture of the AUV carrier changes when the work is unfolded,which affects the dynamic performance of the AUV carrier-manipulator system,which in turn affects the working effect of the AUV.Therefore,it is of great research significance and practical value to study the dynamic performance of the AUV carrier-manipulator system and the control method of the robot attitude to study the operation quality and efficiency of the AUV carrier-manipulator system.In order to study the influence law and control method of the working AUV attitude by the manipulator movement,this paper developed the AUV carrier-manipulator experimental prototype.Based on the small self-propelled underwater vehicle carrier "BeaverII" in the laboratory,a small double joint underwater electric manipulator was developed.According to the characteristics of open frame and small size of "Beaver II" carrier,the technical index of the manipulator is proposed.The selection of joint drive motor and the required power calculation are carried out to complete the structural design of the two joints of the manipulator and complete the joint between the joint and the "Beaver II" carrier.The structural design of the connection module and the joint module between the joints;the selection of the materials of the components of the robot,the selection of the angle sensor for detecting the angle of rotation of the joint of the manipulator,the design of the connection between the sensor and the robot,and the method of sealing the sensor;Installation and commissioning,the "Beaver II" carrier after the robot is mounted,the water is floated and leveled.The dynamic performance of the "Beaver II" carrier after the manipulator was studied,and the kinetic model of the two degrees of freedom of the pitch and the longitudinal direction was established.Aiming at the problem that the manipulator poses the change of the attitude of the carrier,the stress of "Beaver II" after the manipulator is analyzed.The least square method and BP neural network learning method are used to identify the unknown parameters of the longitudinal and longitudinal dynamic models through the pool experiment.In order to verify the accuracy of the established kinetic model,the pool experiment was verified by the "Beaver II" carrier equipped with a robot.The "Beaver II" carrier equipped with a robot is used to control the attitude of the robot during operation.Aiming at the nonlinear and strong coupling characteristics of the operating underwater robot system,this paper designs an adaptive neural network control method.In this method,in order to reduce the influence of model uncertainties on the control accuracy of the carrier AUV,based on the adaptive neural network The method is used to approximate the uncertain factors in the dynamic model,and the stability of the system under the controller is analyzed by Lyapunov theory.The simulation experiment of the AUV hull's pitch attitude and vertical and horizontal composite attitude control is carried out by different angles of the manipulator.To verify the effectiveness of the adaptive neural network control method designed in this paper.The experimental study of the "Beaver II" carrier posture retention problem during the operation of the robot was carried out.In order to verify the practical effect of the adaptive neural network control method designed in this paper,this paper takes the "Beaver II" equipped with manipulator as the experimental object,and conducts multiple sets of pool attitude control experiments,and compares it with the typical sliding mode control method.The multi-set pool attitude control experiments are as follows: 1)When the manipulator moves in step and sinusoidal trajectory,the AUV trim and heel are controlled separately for pool experiments to verify that the method is free for AUV when the manipulator moves in different regular motions.The control effect of degree attitude;2)When the manipulator moves with step and sinusoidal trajectory,the pool experiment of AUV vertical and horizontal composite attitude control is carried out to verify the control effect of the method on AUV composite attitude;3)To verify the method against external The ability of interference,this paper simulates the trim,heel single degree of freedom attitude control and the vertical and horizontal composite attitude control pool experiment under simulated sea current disturbance.