Design And Implementation Of Multi-target Tracking System

Target tracking theory was first proposed in the 1950 s.It is widely used in the fields of precision guidance,battlefield surveillance,navigation and traffic control.Tracking system is significant for specific airspace scenarios such as airports.It not only watches the airspace in real time,but also provides data support for regional track analysis and identification of friend or foe.This thesis analyzes the specific needs and data characteristics of the multitarget tracking.It designs and carries out a matching multitarget tracking system.The design of the system adopts a hierarchical architecture model.The system first receives radar data,classifies and calculates the data.It uses time difference location algorithm to gets target position information,and then performs data association filtering.It calculates target tracks and performs human-computer interaction through a visual interface to oversee and record the target track in real time.This system mainly contains the following parts: 1.Data processing.This module communicates with the base station server.It gets the target track according to the data preprocessing method and the data association method.Then it sends the calculated result to the visualization platform.2.Real-time monitoring.By querying the results of data processing in real time,flight situation and target information in the monitoring area are displayed in the front end.The system supports setting focused target.3.Query historical tracks.It supports query historical information by time or target number.4.Map service.The system supports setting of geographic labeling information such as airports,alarm zones.It also provides plug-ins such as map zoom,latitude and longitude labeling,and scale to ensure a good user experience.The system adopts the J2 EE architecture as the overall technical solution.It uses B/S structure and realizes a front-end human-computer interface.Besides,it uses My SQL to persist the data.The key technology of a tracking system is track calculation.To solve the practical engineering problems,this system uses a mature track initiation method and data association algorithm to calculate the track.Then,this thesis makes improvements according to the actual problems and data characteristics.It mainly solves two problems in track calculation.The first is the repeated track initiation caused by unstable data frequency.This thesis analyzes the received message data,classifies the data,and uses the density-based DBSCAN algorithm to select an appropriate threshold for clustering.Simulation showed the number of repeated start tracks was significantly reduced.The second problem is the relatively large track error after filtering.We use a weighted data fusion method to get the final track by using the high-precision analytical data of the passive radar.Simulation showed the fusing track has smaller relative mean square error.The system has been accomplished in all the coding and testing work,as well as related hardware equipment.Test results in the field are good.It meets the design needs.