Study On Target Tracking Technology Based On Radar Network Data Fusion

With the continuous development of computer technology,communication technology and microelectronics technology,and the increasing complexity of modern warfare,single radars cannot meet the needs of combat systems.Various types of Radar Network for complex application backgrounds have appeared in large numbers and are extensive used in Electronic Warfare.Compared with single radar,Radar Network greatly expands the coverage in space and frequency domains,and enhances the radar's comprehensive detection capability through data fusion technology.With the continuous innovation of data fusion technology,networked radars can integrate radar data more efficiently.Faced with a large amount of data,people urgently need to study more effective target tracking algorithms to process the data to meet the real-time and accuracy of the tracking system.In the context of this application,based on the Interactive Multi-Model algorithm,the main content of this thesis is studying the maneuvering features such as target heading information,to improve the state transition probability matrix of the interactive multi-model algorithm,accelerate the running speed of the algorithm and increase the tracking system accuracy.The main work of this thesis is as follows:First of all,this thesis starts with an analysis of the main problems faced by the Radar Network.Although the development of the Radar Network compensates for the lack of single radar detection information and the single detection direction,it also brings about some problems,such as the configuration and deployment of the Radar Network,intranet radar communication.,intranet radar space-time registration and data fusion processing.Focusing on the large amount of data detected by multiple radars in the fusion center,the characteristics of the centralized and centralized data fusion technology used in the fusion center are discussed,and the data fusion problem when the Radar Network is used to coordinate detection of aircraft targets in space,time and frequency.And so on,and a gradual study of the process of convergence of Centralized Radar Network spot tracking,as much as possible to increase the trace accuracy for subsequent data processing.Then,this work focus on the tracking filter method.The most commonly used is a single model tracking filter Method,such as Constant Velocity Model filter,Constant Acceleration Model filter and Singer model filter.But in actual projects,the state of motion of the target is often very complex,usually a random combination of uniform speed,uniform acceleration,circular motion,subduction and pull-up motion models,and even a sudden change in the target state,it will lead to a strong maneuver of the target.If tracking filter continues to use a single model filter,it will tend to result in a loss of tracking accuracy,even with missing targets or wrong tracks.In order to solve this problem,based on the Interactive Multi-Model Algorithm,the parameter values such as instantaneous heading,speed and acceleration are calculated based on the position,orientation and other information in the point data,and the model transfer function is optimized accordingly.The accuracy of algorithm tracking is improved under the conditions of model computation,and it is more effectively applied in engineering practice.Finally,this thesis focuses on spot tracking fusion and target tracking for the simulated centralized network radar spot tracking data.This work compares and analyzes the processing results of various tracking filtering algorithms.It is proved that this method improves the target tracking accuracy and speeds up the calculation speed.It achieved the desired effect of this thesis.