SPECSweb post-tracking classification method

The Specular-Cued Surveillance Web (SPECSweb) multistatic tracker effectively reduces false track rate through the use of two amplitude thresholds. A high threshold is used to identify the occurrence of high-strength specular detection cues for track initiation. A low threshold is used for selective extraction of additional detections for track extension forward and backward in time. This approach significantly reduces the node-to-fusion-center communication loading requirements as well as reducing the output false track rate. At the time of the specular-cued track initiation, a historical (back)-track is reconstructed and output. In some cases, there may still be false tracks at the tracker output. We demonstrate a post-tracking classification method which identifies and removes any residual false tracks. Classification feature information for track-associated contacts is passed through the tracker and is available at its output. Discriminating true target tracks from false ones at the tracker output is made using a sequential probability ratio test (SPRT) based on feature information and kinematic track-quality. The method is shown to be effective in reducing false tracks when applied to a simulated data set.