| The demand for underactuated unmanned surface vehicles(USVs)for maritime operations is increasing,and the study of guidance control strategies in complex marine environments is crucial for underactuated USVs to achieve accurate target tracking.The USV for ocean target tracking need to cope with the challenges of wind and wave interference at sea,sensor noise and obstacles in the working area.In this thesis,we address these problems by using the dynamic A* algorithm to generate feasible waypoints to find the Dubins curve smoothing path and study the vector field-based guidance laws to improve the ability of USVs to accurately track the desired path and meet the docking constraints of USVs and autonomous underwater vehicles(AUVs).The main contributions of this thesis are as follows:(1)Aiming at the navigational path planning problem of dynamic target tracking and docking of USV,grid detection and feasible domain search are implemented for the unknown region,and Dubins path smoothing method based on A* algorithm is proposed.In the thesis,the detection of unknown obstacles is achieved by using the Lidar on the USV,and multi-sensor fusion is realized with the GNSS/IMU on board the USV and AUV to obtain the vehicle position and yaw angle.The information obtained by the sensors is passed to Cartographer algorithm node to implement localization and mapping,and to complete A* algorithm path planning for dynamic scenes to find the optimal path around the obstacles.To avoid oscillations and curvatures during waypoint tracking,Dubins curve smoothing optimization is implemented for discrete waypoints.When the USV reaches the docking start point directly behind the AUV,the USV uses the AUV as the target point for trailing straightline tracking,and dynamically adjusts the position and orientation of the USV to avoid collision with the AUV during docking.(2)Focusing on the problem of low accuracy of patrol line brought by waves and wind during USV following path,this thesis proposes a fusion strategy based on vector field and integral line-ofsight(LOS)guidance control law to improve the guidance effect and reduce the influence of constant external disturbance.The fusion guidance strategy is divided into two cases,linear following and curved following,introducing the smoothing parameter in linear following driving and vector field in turning to speed up the convergence to the desired path.To further reduce the effect of time-varying external disturbances,the vector field-based adaptive LOS guidance law is proposed to reduce the oscillation of the desired path tracking with time-varying lookahead distance and turn vector field prediction.The proof of κ-exponential stable convergence for curvature-variable desired path following is given in the thesis,and the effectiveness of the controller is experimentally tested.In this thesis,the controller nodes of USV and AUV are built sequentially based on the ROS system,and finally the control performance simulation and hardware experiments are performed on Gazebo and Software In The Loop(SITL)to verify the effectiveness and robustness of the proposed guidance law and to realize the target tracking of underactuated USV.This thesis mainly shows the theoretical research on USV target tracking and docking,and the proposed strategy can be used for the control of many kinds of marine vehicles,which can also provide technical and theoretical support for subsequent experiments of unmanned vehicles. |
