An adaptive-gain alpha-beta tracker combined with circular prediction for maneuvering target tracking

Abstract

iln atlaptive-gain 0-9 tracker cornbined with circular prediction for maneuvering target tracking is proposed as a nonlinear prediction filer. For track-while-scan systems using phased array antenna, the reduction of prediction error is ill1 very important issutr. The purpose of this scheme ir t o achieve good noise reduction and get better maneu-\ c:r-following capability. Simulation results for several targci profilcs arc. incliided for a comparison of the perfortniiiices of our proposed scheme with that of conventional I r i d i t i g filters.

A novel manuevering target tracking algorithm based on moving slide window

“…The coefficient α for the α-β tracker is obtained by the criterion based on the best linear track fitted to radar data in a least squares sense [8]. The coefficient β for the α-β tracker is used 0.05 and the other parameters are obtained by the criterion based on the best linear track fitted to radar data in a least squares sense [4]. To verify the proposed algorithm, we assume target moves every 0. x represents true position.…”

“…The well known α-β tracking filter [1,2] and Kalman filter [3] estimate position and velocity by using measurements of maneuvering targets. The Kalman filter shows excellent tracking performance, if the target dynamics and statistical characteristics of measurement noise are available, which is difficult to obtain in advance [4].…”

“…Despite of simplicity, the tracking accuracy of the α-β tracker is not guaranteed and sometimes target is lost when the target trajectory is nonlinear in noisy environment [4]. Also, the performance of the α-β tracker heavily depends on suitable choice of α and β, which make it difficult to be used for maneuvering target with AWGN (Additive White Gaussian Noise) [5].…”

Maritime radar simulator based on DSP board using switched slide window tracker

“…However, it is difficult to find the statistical characteristics of the maneuvering target in advance. Furthermore, the Kalman filter has a heavy computational cost [3].…”

“…Especially the α-β tracker is important how to set α and β. If α and β are not suitable, when a target is moving, the predicted target could be lost because of the effect of the increasing linear prediction error to the nonlinear prediction [3].…”