A Measurement Correlation Approach for Multitarget Tracking in a Noisy Environment

Abstract

In multitarget tracking, measurement correlation uncertainty occurs when remote sensors, such as radars, yield measurements whose origin is uncertain. Using incorrect measurements in multitarget tracking systems leads to tracks loss. In such cases, efficient measurement correlation methods are needed to select measurements from many to be used to update the target tracks of interest in the tracking systems. This paper proposes a measurement correlation approach for multitarget tracking in a noisy environment. In this approach, measurements-to-targets correlation is computed across all targets and measurements based on minimization of weighted total squared errors. For a given track, the measurement that has the maximum correlation is used for updating the target track. The proposed correlation approach is applied to a scenario of multitarget tracking system and performance comparison with other correlation approaches is also presented. The results showed that performance improvement in terms of correct measurements correlation is achieved.