Fast Tracking Method. Short-range Radar Detection System

In this dissertation the key problems of accurately tracking high speed maneuvering targets are systematically investigated in a short range multi-static radar system. Strenuous computation periods and intensive real-time requirement are typical characteristics of short range tracking system. Effectual research of fast tracking methods will facilitate the comprehensive application and farther advancement of this technology.To investigate the problem of state estimation with intensive real-time requirement in short range multi-static radar detection system, a parallel extended Kalman filtering algorithm (PEKF) is derived. When constructing the system's measuring model, the acceleration vector is introduced into target motion equation. By introducing the way of parallel calculation, the multidimensional iterative filtering course is separated into two parallel 3-D sub-courses to estimate the target's position and velocity. To insure real time synchronization, the predicted value of current position is used as input to estimate the target's current velocity. Computer simulation results show that the algorithm can accurately calculate the target's position and velocity with fast speed even if its initial values are worse. Meanwhile, the algorithm has good convergence and stability.To investigate the problem of improving detection performance in large airspace for multi-static radar net, a variable observation adaptive tracking algorithm (VOAT) based on sensors distribution optimization according as airspace segmentation scheme is derived. During the computational process of target tracking, the sub-airspace of the target is calculated by its position estimation and observation distribution matrix is introduced to adaptively adjust measurement system of equations. Low precision observing equations are eliminated and high precision observing equations are reserved. Simulation results show that VOAT not only ensures the precision of state estimation but also reduces the dimension of observation equations. The accuracy and real-time capability of maneuvering target tracking are improved.To investigate the problem of multi-target tracks initiation in clutter for T-R~S multi-static radar system, a fast initial traces calculation algorithm syncretized with observing distance-sum and Doppler velocity is derived. Measurement association preprocessing module and tracking gates are used to eliminate redundant candidate associations. The S-D assignment algorithm is adopted to search minimum cost measurement associations and calculate initial traces. Each target's stable trace is estimated by PEKF. Courses of multi-target tracking and S-D assignment are parallelized to improve real-time performance. Simulation results show that the algorithm not only accurately tracks initial multi-target states with fast speed, but also has good stability and convergence. It's appropriate to calculate initial traces of multiple short range high speed maneuvering targets.To investigate the problem of measurement-trace association for short range multi-static radar system, a fast tracking algorithm based on parallel processing mechanism is derived to track multiple high speed maneuvering targets in clutter. The whole measurement association course was divided into some parallel sub-courses, moreover, PEKF is adopted to execute iterative filtering processes with synthetic distance-sum and Doppler velocity. Computational course of each established target's conditional probability is divided into N (number of receivers) parallel sub-courses and can be applied to multi-target circumstance. Simulation results show that the algorithm can accurately track multiple maneuvering targets in clutter with fast speed. Meanwhile, it has good convergence and stability. The parallel processing mechanism improves its real-time performance and makes it fit for accurately tracking multiple short range high speed maneuvering targets.