Methods Of Multi-sonar Fusion Detection And Tracking

Due to the deepening of the underwater warfare, single base sonar can't meet the needs of underwater battlefield. Because of the far detecting target distance, good detection performance and the system strong ability to resist destruction, multi-base sonar system has become the development trend in the current study. In this paper, multi-sonar fusion detection and tracking technology are researched. And through Songhua lake trial data based on multi-base active sonar detection system platform, multi-base sonar distributed fusion tracking algorithm and centralized fusion tracking algorithm are verified.Firstly, based on a single sonar signal detection, under certain sentencing guidelines of stations it studies the hard fusion detection algorithm, soft decision fusion detection algorithm and K/N fusion detection algorithm commonly used in actual engineering. And in the case of a single station using cell averaging CFAR detector, it gives receiver operating characteristic curve of the three fusion detection algorithm and analyzes multi sonar system target detection range at an equal probability of detection of each algorithm.Then based on a single base station using Kalman filter tracking algorithm the paper studies track association technique and track fusion technology in multi-sonar distributed tracking algorithm. And it studies weighted method, sequential method, and track association algorithm based on the classic allocation method. In bistatic multi-objective situation each algorithm are simulated and compared. And it shows the performance effect of different measurement error for each algorithm. The error curves of track fusion technology at different measurement noise are given. Multi-sonar distributed tracking algorithm is verified through lake trial data.Finally on the foundation of joint probability data association algorithm ( JPDA ), it studies multi-sensor joint probabilistic data association algorithm ( MSJPDA ) of the sequential processing structure. And in bistatic single and double objective situation it simulates and gives the performance effect of the clutter number, process noise,measurement error. Through the trial data multi-base sonar distributed fusion tracking algorithm and centralized fusion tracking algorithm are verified.