Real Time Multi Sensor Multi Target Data Association Over Sensor Network

Abstract

Data association is an integral process in tracking multiple targets with multiple sensors in a cluttered environment. With the help of data association, we obtain the relationship between sensor measurements and existing tracks. Tracking moving targets from a stationary position is likely a potential problem for losing tracks or track mingling. To overcome this problem, we employ a clustering algorithm solution developed using the nearest neighbour technique with the help of an Extended Kalman Filter. This paper focuses on finding out the real-time clustering solution for each target, having at most one measurement from any heterogeneous sensors at a particular timestamp. Here we employ a novel multi-target tracking algorithm using three basic association techniques, i.e. track-to-track association, measurement-to-measurement association and track-to-measurement association. Our real-time data simulation experiments show that the data association algorithm implemented works effectively and gives a good efficiency in real-time multi-target multi-sensor environ-ments.