Research On Moving Target Tracking In Shallow Sea Of Near-shore For UUV

Target tracking is the crucial technology of underwater unmanned vehicle (UUV) to achieve submarine pipeline maintenance, harbor defense, intelligence gathering, underwater recovery and other missions, which has a wide application foreground. In order to get more practical solution of the target pursuit problem in submarine horizontal plane, the dynamic planning result is taken as the UUV motion control objective, which is quite different from trajectory tracking or path following. This paper mainly discusses the following aspects:Firstly, we establish the tracking system model. In the inertial coordinate system and the moving coordinate system, the kinematic model, dynamic model and its formulation under constant current condition of UUV are presented based on the analysis of motion rules and mechanical characteristics. Three models of moving targets under the conditions of constant velocity (CV),constant acceleration (CA),constant turn rate (CT) and two sensor observation models are given. And furthermore, simulations are designed to verify the validation of all the models above.Secondly, we study state estimation method based on target model and adaptive filtering for UUV data. On account of the basic principle of the Kalman filter, discrete Kalman filtering (DKF), unscented Kalman filtering (UKF) and interacting multiple Model (IMM) are designed individually to get the unknown parameters and state estimation. Then,simulations of target state estimation and result analysis are given to prove the three algorithms. On the other hand, grey dynamic prediction model is applied for data filtering with the advantages of smoothness and real-time in view of poor information and small samples. Improvement of adaptive weighting fusion between estimate and sampling value is made to increase the outlier removing ability.Thirdly, considering the control goal of UUV navigation safety and tracking accuracy ,moving target tracking strategy based on sight switching is developed. What's more,with the constraints of the velocity and yawing angular velocity, the proposed guidance strategy relies on the relative motion model between UUV and target applies for both static and moving obstacle environment.Finally,a non-singular terminal sliding mode algorithm has been brought to design the vehicle velocity and yawing controllers. The stability of the controller is verified by the theory of Lyapunov. In the end, system simulations for target tracking base on state estimation, dynamic planning and motion control is designed, showing the feasibility and applicability of the program through the rotational moving target, CV-CT maneuvering target in static obstacle environment and moving obstacle environment.