Multistatic tracking of dim targets in large sonobuoy fields using the metron blind data set

This paper describes an analysis of simulated multistatic active sonar data from a large sonobuoy field having large measurement errors, a low probability of detection and a high false alarm rate. The data comprises the first of five scenarios collectively known as scenario one of the Metron Blind data set. Single- and multi-hypothesis Kalman filter trackers are used to process contact measurements from the field, using parameters calculated from an initial analysis of the data. Track management is implemented using an m-out-of-n criterion and then a sequential probability ratio test (SPRT). Gaussian mixtures are used to accommodate the transformation of measurement errors from sensor coordinates into tracker coordinates. Results are presented for each tracker configuration.