International Journal of Advances in Computer Science and Its Applications
Author(s) : HYUNG JUNE KIM, TAE HAN KIM, TAEK LYUL SONG
This paper presents a smoothing data association algorithm for a single target tracking in clutter. The proposed algorithm fuses the forward estimates and all the available measurement information retrodictions (but not backward track estimates) within the smoothing window to obtain the smoothed estimates. The measurement information retrodictions are obtained using the one-step-backward information filter propagation for fast calculation. The simulation studies show that the proposed algorithm improves the false track discrimination performance with similar root mean squared errors as the existing smoothing integrated probabilistic data association algorithm.