Optimization Of Joint Detection And Tracking Algorithm Based On Sequential Detection

In weak target environment, the detection-tracking-integration algorithm uses allobservation data without a threshold to estimate a track which avoid high probabilityof miss for traditional algorithms that led significant loss of target information. Anddetection and tracking processes are coupled that detection parameters are adjustedwith the target position feedback by tracking, which improve the probability ofdetection in the case of weak target. At last, it helps improve a joint performance ofdetection and tracking. Because of coupling detection and tracking processes, thestructure of detection-tracking-integration algorithm is very complex and nonlinearthat is difficult for quantitative analysis of the performance. Present study focuses onthe research and improvement of algorithms based on different criteria.Joint detection and tracking algorithm based on sequential detection is one way ofdetection-tracking-integration algorithm. The paper is focused on the optimization for ajoint detection and tracking algorithm based on sequential detection in early warningradar environment. And the classification optimal algorithm which the accumulatednumber of frames is fixed is seen as reference algorithm in optimization analysis,whose performance is also needed a detail quantitative analysis. In the paper, the towalgorithms’ performance prediction, establishment of the quantitative relationshipbetween parameters and performance and solving optimal parameters of joint detectionand tracking algorithm are focused analyzed. The main contents are as follows:1. The probability density function (PDF) of the classification optimal algorithmand joint detection and tracking algorithm is analyzed to predict the values of thedetection probability, tracking probability and the average number of accumulatedframes, which are good matched with real Monte Carlo results.2. The prediction results of the two algorithms are used as sample data to regressanalyze the quantitative relationship among the algorithm parameters, performance andthe average number of accumulated frames, which cut down the estimated error. At thesame time, the effects of the parameters on the system are studied.3. Focused on the quantitative relationship between the algorithm parameters and the performance, two optimal strategies are provided for optimization of the detectionand tracking algorithm joint performance. One is to find the best performance underthe constraint of the max average accumulated frames, the other is to find theminimum average accumulated frames under the constraint of the joint performancewhich is must beyond the one of traditional track-before-detect algorithm. Withoptimization, the values of parameters which meet the optimal strategy are predictedthat achieve joint detection and tracking optimization algorithm.