Research On AUV Path Following Method Based On Model Predictive Control

The competition between countries for ocean development and resource is becoming more and more intense,Autonomous Underwater Vehicle(AUV)has attracted more and more attention because of its good security,strong autonomy,fast search speed,strong maneuverability and High modularity.Compared with Remotely Operated Vehicle,AUV has stronger autonomous capability,but at the same time it brings higher requirements for control performance.Therefore,it is necessary to study more advanced control methods to meet the complex and high-precision operation requirements.The ability to track a given path accurately is one of the most important index to measure the operation capability of an AUV.In most of the tasks performed by AUV,path following task are included.In this paper,the plane path tracking control method of AUV is studied.Firstly,a path following controller is designed based on the theory of Model Predictive Control(MPC).The proposed controller is applied to different paths and different initial points,and the simulation results are compared to verify the effectiveness of the proposed control algorithm.Considering the uncertainty of AUV model parameters,a stochastic MPC method is proposed.By using polynomial chaotic expansion approximation theory,the direct mapping relationship between uncertain parameters and output response is established.The proposed stochastic MPC method is compared with the control method without considering the uncertainties of model parameters through simulation experiments,and the robustness of the proposed control method to parameter perturbation is verified.Finally,a Lyapunov-MPC(LMPC)control method based on Control Lyapunov Function(CLF)is proposed.This method combines the characteristics of MPC and CLF,which makes the controller not only take small gain coefficient to ensure sufficiently large attraction area,but also obtain better control effect by rolling optimization method.Finally,through the simulation experiment of path following under external disturbance and model parameter perturbation,the effectiveness and robustness of the proposed algorithm are verified by comparing the LMPC controller with the traditional back-stepping controller.