Incorporating track uncertainty into the OSPA metric

This paper proposes the use of the Hellinger distance in evaluating the localisation error in the OSPA metric. The Hellinger distance provides a measure of the difference between two distributions and is used here to measure the difference between the true and estimated targets where the true and estimated single-target states are characterised by Gaussian distributions. The OSPA metric is used to evaluate the performance of several multisensor PHD and CPHD filters. The importance of introducing track covariance into the metric is demonstrated through the application on several multisensor PHD and CPHD filters. In particular, we are able to identify filters that provide a poor estimate of the uncertainty associated with each track.