Application Of Multi-sensor Based Trajectory Prediction In Target Tracking

Target tracking is widely applied in military and civil fields,with the rapid development of modern aerospace technology,the accuracy of target tracking trajectory prediction is also increasing.Multi-sensor-based target tracking is a typical example of data fusion technology,and its tracking performance is superior to a single sensor.Among them,the research of track extraction and trajectory fusion algorithm for multi-sensor has important practical significance in target tracking and prediction.Based on the analysis of the existing target tracking algorithms,this paper focuses on the target trajectory extraction prediction and trajectory fusion prediction techniques.Aiming at the shortcomings of existing algorithms,an Interacting Multiple Models-Adaptive Filter Algorithm and a Fusion Algorithm base on Separate the Independent Variance are proposed,to improve the trajectory prediction accuracy,the results obtained in this paper include the following aspects:First of all,in view of the problem that the current target motion model usually does not match with the actual motion state and the tracking accuracy is low,an Interacting Multiple Models-Adaptive Filter Algorithm is proposed.The algorithm can solve the problem of large trajectory extraction error due to the single target filtering model by establishing a variety of different motion states and filtering models of the target.At the same time,aiming at the problem that the model transition probability matrix is fixed in the existing algorithms,a real-time adjustment model of Markov Transition Probability Matrix is proposed from the perspective of practical application,so that the probability of multi-model transition probability can be improved with the measurement information of the target and updated in real time to improve the trajectory extraction accuracy and trajectory prediction accuracy in the target tracking system.Secondly,aiming at the problem that the existing fusion algorithm introducedrelevant errors in data preprocessing,a Fusion Algorithm base on Separate the Independent Variance is proposed.The algorithm can solve the problem of large computational complexity of mutual covariance and large error of data processing,so as to improve the precision of trajectory fusion prediction.Finally,the target tracking system based on radar is modeled as much as possible by using System Vue platform to transmit the received signal mechanism and simulate the real environment's speed,noise and other interference signals.This paper presents an interactive multi-model adaptive filtering algorithm based on Matrix transition probability real-time adjustment model and the trajectory fusion algorithm with separate independent variance are simulated and tested.The simulation results show that compared with the existing algorithms,the proposed algorithm can significantly improve the prediction accuracy of trajectory extraction and trajectory fusion in the target tracking system model.The research results in this paper have certain application value in the field of accurate target tracking and prediction.