Application of UK-GMPHDF algorithm based on IMM in multiple maneuvering targets tracking

Yan-ling HAO,Fan-bin MENG,Feng SUN,Feng SHEN

2011-01-01

Abstract:For the purpose of solving the problem of nonlinear filtering and the multiple maneuvering targets tracking, a novel unscented Kalman implementations of the Gaussian mixture probability hypothesis density filter (UK-GMPHDF) based on interacting multiple model (IMM) is presented. The adaptive ability to various target maneuvering models is combined with the advantage of higher accuracy and lower computation load provided by UK-GMPHDF. Furthermore, the UK-GMPHDF avoids the data association problem and is able to jointly estimate the time-varying number and their states. Under nonlinear system and clutter environment, the proposed algorithm is compared with constant turn (CT) model and current statistical (CS) model based on UK-GMPHDF in maneuvering target tracking. The experimental results show that this novel algorithm can significantly improve the tracking performance and reduce the multi- target miss-distance.