Trajectory Tracking Control Of An Underwater Swimming Manipulator With Energy Optimization And Fault Tolerance

The underwater swimming manipulator is a new robot composed of an underwater bionic snake-like robot and several propellers.Because of its ability to move flexibly and autonomously,its application prospects are extensive.However,the underwater swimming manipulator system has the characteristics of strong coupling and high uncertainty,which brings challenges to its tracking control.Moreover,because the underwater swimming manipulator needs to work for a long time in the marine environment,it is of great significance to optimize energy consumption to improve the endurance of the robot.The over-actuated and high redundancy of the underwater swimming manipulator makes it possible to improve fault tolerance.This paper uses the underwater swimming manipulator as the research object and proposes trajectory tracking control with energy optimization and fault tolerance.The main research contents are as follows:First of all,a mathematical model is introduced for the new type of underwater swimming manipulator system.Based on the Lie group,Lie algebra,spinor,etc.,the kinematics model of the underwater swimming manipulator is established.According to the characteristics of the high dimensional state variables of the underwater mobile manipulator’s multi-link series structure,the underwater manipulator’s dynamic model is established based on the Lagrangian equation,which avoids the complexity of the Newton-Euler method in the high-dimensional system.Then,with energy optimization and precise control as the target,considering the influence of environmental disturbance and model uncertainty,the trajectory tracking control of the underwater swimming manipulator is carried out.In this paper,the inner and outer loop cascade control architecture is adopted.The outer loop controller aims to minimize the restoring torque and increase tracking accuracy while considering joint constraints and configuration singularity avoidance.The inverse kinematics controller is described as a multi-objective optimization problem,and the solution is obtained based on the dynamic recurrent neural network method.The expected trajectory of the generalized coordinate of the underwater swimming manipulator is thus generated.For the inner loop controller,this paper gives a dynamic controller based on the Barrier Lyapunov function and backstepping method.The torque allocation then generates the actual control variable.The radial basis functions neural network is introduced to estimate and compensate the disturbance and the unknown part of the model in realtime to solve the problem of external disturbance and model uncertainty.Through simulation verification,the proposed algorithm guarantees the accuracy of end effector trajectory tracking and significantly reduces the energy consumption during the movement.Finally,a fault-tolerant control system for the underwater swimming manipulator is designed for two types of faults: free swing and locked-up of a single joint.This paper establishes a mathematical model of the underwater swimming manipulator under the condition of joint failure and analyzes the impact of the failure on the controllability of the robot.In the case of freely-swinging joints,based on a multi-task priority controller in the operating space,the singularity avoidance task of the actuator is added as a strong constraint to the controller design to ensure the controllability of the underwater manipulator.In the case of the locked joint,based on those mentioned inner and outer loop cascade architecture,combined with the torque redistribution strategy,a faulttolerant controller is designed.The simulation results show the effectiveness of the proposed algorithm for fault-tolerant control.