Multistatic false alarm mitigation using a contact proximity feature Analysis and MHT processing of the Metron data set

The Metron data set was developed to facilitate and encourage collaborative analysis and comparisons of tracking methodologies within the Multistatic Tracking Working Group (MSTWG) of the International Society of Information Fusion (ISIF). It was designed to be very challenging, with dim targets and large bearing and range errors in a high noise multistatic environment. Tracking results to date have shown limited success with the full data set, although some groups have shown good results using only the tagged target contacts. This paper describes a viable method for detecting and tracking targets in partial clutter using a Kalman filter-based multihypothesis tracker. The use of Gaussian mixtures is discussed and demonstrated to address the large bearing and range errors. The high false alarm rate is mitigated using a contact proximity feature that makes use of the sequential probability ratio test-based track scoring of a multihypothesis tracker (MHT) and avoids some of the challenges and limitations of contact clustering prior to tracking.