Joint nearest neighbor data association based on interacting multiple model Kalman filtering

As one of the key issue in multiple target tracking, track association is very important and difficult, especially in a heavily multi-corssed environment. In the long-range multitarget tracking engineering practice, relative to the clutter problem, data association for target crossing is the key problem. There are several data association algorithm, such as NNDA and JPDA, but the accuracy of NNDA is too low and the speed of JPDA is too slow. Combined with NN and JPDA ideas, this paper propose the joint nearest neighbor data association(JNNDA). This paper introduce the process of JNNDA and compare the tracking error conditions of NNDA and JNNDA, proved that the accuracy of JNNDA is high than NNDA. The simulation results shows that the accuracy of JNNDA reach the level of JPDA and the speed of JNNDA reach of the level of NNDA.