Research On Non-cooperative Targets Position Estimation And Tracking Based On Underwater Sensor Networks

Target tracking based on underwater sensor networks inherits the characteristics of small nodes,low power consumption and flexible distribution in the sensor network,which can provide more real-time and effective data support for underwater non-cooperative target tracking.The environmental characteristics and target characteristics of underwater noncooperative target tracking are studied in this paper.For the purpose of more reliable tracking,a tracking algorithm based on group estimation strategy and a tracking algorithm based on resource scheduling strategy are designed to fully explore the advantages of distributed collaborative decision.The work of this paper mainly includes the following two aspects:(1)To overcome the hysteresis of underwater acoustic propagation in underwater positioning and the impact Non Line of Sight(NLOS)errors,an asynchronous distances measurement method and a tracking algorithm with NLOS measurements based on group strategy are proposed in this paper.In the measurement method,the asynchronous positioning method is used to acquire the distancs,which weakens the adverse effect of propagation delay.In the positioning algorithm,the position accuracy decline caused by NLOS measurement noise is effectively reduced through multiple position estimations.Experimental simulation results show that the algorithm can improve positioning accuracy for underwater tracking with NLOS measurements.(2)Aiming at the problems of measurement mismatch,false alarm interference and uncertainty of target number caused by non-cooperative target characteristics,a particle filter target tracking algorithm based on resource scheduling strategy is proposed.Through cooperation between nodes,certain information such as the number of targets and the initial state of the targets are provided,thereby avoiding measurement mismatch and false alarm to reduce positioning accuracy and achieve high-performance tracking.Considering the multi-target intersection and multi-target approaching scenarios,a recursively updated model is designed to track multiple targets in real time while reducing false correlation results.Experimental showes that the designed method can achieve high-performance tracking results when intersections and approaching scenarios arising.