We developed algorithms that allow us to detect an unknown number of targets that appear and disappear at unknown points in time. In contrast to the traditional detection methods (Wald/Neyman-Pearson), the problem of detecting a target with unknown moment of appearance is formulated and solved as a quickest detection problem (change-point detection problem). This problem involves optimization of a trade-off between the detection delay and the false alarm rate. The developed sequential algorithm detects a target with as small average delay as possible (after it appears) with a fixed frequency of false alarms. In addition, the algorithm detects target disappearance, also with the smallest delay, which makes it possible to almost immediately interrupt tracking when a target disappears. The decision statistics use the results of track-before-detect, i.e. the estimates of a target's spatial location based on the optimal spatial-temporal nonlinear filtering.
Click here to watch Detection and Tracking in real time. This movie shows detection and tracking in action with the use of the shipboard IRST data (courtesy of the SPAWAR Systems Center, San Diego, CA). There are two targets in the scene that appear and disappear at unknown points in time. Both events, target appearance and target disappearance, are detected with a relatively small delay. The signal-to-noise ratio in these experiments was extremely small (-3dB to –6dB).