DISTRIBUTED SEQUENTIAL NEAREST-NEIGHBOR MULTITARGET TRACKING ALGORITHM

An efficient distributed multitarget tracking algorithm is proposed. T his distributed tracker consists of two main components: local sensor level trackers and a track fuser. In the track fuser, local tracks fro m sensors are first transformed to a common co-ordinate system, and sy nchronised by a linear Kalman filter. A track correlation technique ca lled sequential minimum normalised distance nearest neighbour (SMNDNN) method with the majority decision making (MDM) logic is used to corre late tracks from different sensors. The correlated tracks are fused us ing a sequential minimum mean square error (MMSE) fusion approach. The SMNDNN correlation converts the multisensor track correlation problem to a one-to-one nearest neighbour assignment, and the sequential MMSE fuser with the MDM logic combines the tracks optimally if the majorit y of sensors report similar tracks. Simulated data under various track ing conditions are used to evaluate the feasibility and effectiveness of this proposed distributed tracker.