Analysis of the cascade of track-before-detect and track-after-detect tracking algorithms

Advanced track-after-detect (TAD) trackers are able to operate with detection thresholds as low as 9.5dB with the use of track features. At lower thresholds the increased number of false alarms inhibits track confirmation. In order to track weaker targets, the target SNR must be increased prior to detection. Assuming that the SNR has been increased as much as possible through signal processing, further increases in SNR can be obtained by preceding the detection threshold with a track-before-detect (TBD) algorithm. This paper analyzes the performance of the cascade of a TBD and a TAD tracking algorithm. The updates provided to the TAD tracker must be independent. To achieve this independence, the TBD tracker must effectively operate on non-overlapping segments of data. The longer the data segment the greater the SNR gain provided by the TBD tracker. Since the segments of data are non-overlapping, the update rate into the TAD tracker decreases as the SNR increases. The lower update rate results in a greater increase in the state error covariance of the TAD tracker between updates. The larger state error covariance delays track confirmation and increases the expected time-to-confirm a track. The increased expected time-to-confirm offsets the benefits of the increased SNR provided by the TBD tracker. This paper analyzes the tradeoff between the SNR gain and the expected track confirmation time. Track features provide a conduit around the detection threshold and into the TAD tracker for information pertinent to track confirmation. Track features have been shown to provide a significant improvement in track confirmation time. The performance of track features is also analyzed and compared to the performance of the TBD tracker.