Research On Underwater Multi-target Track Correlation Method Based On Expectation-maximum Clustering

Underwater target trajectory correlation is a technology that uses sensors to obtain the trajectory point set of each underwater target,and then corresponds the trajectory point set to the real target.Trajectory correlation serves as the basis for sensor information fusion.In a complex underwater environment,if sensors are in the same monitoring area and leak detection occurs,random false alarms will cause the target trajectories displayed on multiple sensors to be inconsistent.,So that the trajectory corresponding to the real target may not be found in the trajectory set reported by different sensors,which will make the originally complicated trajectory association problem more difficult.If the target trajectories are accurately correlated in an underwater noise complex environment,it is of great significance to the development and utilization of marine resources.Aiming at the underwater target trajectory association problems such as errors in underwater measurement and inconsistent targets reported by sensors,this paper proposes a Gaussian Mixed Integer Nonlinear Programming(GMP)algorithm,which uses trajectory hierarchical to remove noise,to build a Gaussian mixture model based on the topological information between trajectories,and to construct a novel mixed integer non-linear programming using maximum likelihood to find the matching relationship between underwater target trajectories,and to continuously sensor bias The method of estimation reduces the association bias,introduces the idea of weighting,and obtains the optimal closed solution in Gaussian mixed integer nonlinear programming through the expectation maximization clustering algorithm.During the expectation maximization phase,the correspondence relationship of the trajectories is obtained,and finally the underwater target is obtained.Track correlation results.The simulation experiment results show that the GMP algorithm proposed in this paper is effective for the GMP algorithm,REP algorithm,FFT algorithm,and FCM algorithm in three-dimensional space under different target numbers,different sensor angle ranging errors,and different sensor detection probability.The average positive correlation rate is compared,that is,the experimental results are averaged by 100 Monte Carlo simulations to obtain the ratio of the number of correctly correlated targets and the number of homologous trajectories to detect the robustness of the algorithm.The results show that the GMP algorithm is in False alarms have good accuracy and robustness in associating underwater multi-target trajectories.