Skip to Main Content
A new track initialization method for remote sensing systems that uses the amplitudes of the measurements to provide some statistical separation between true and false contacts is presented. In the proposed method the data are partitioned into Hough transform bins, and Page's test is performed on the data in each bin. The new method embodies the best ideas of existing sequential and batch methods, exploits both the amplitude and the temporal information in the data, and minimizes the average latency for initializing target tracks for a specified false track initialization rate. Analytic performance predictions for the proposed method for the Swerling I signal model with Poisson clutter are developed and compared with the theoretical performance of a commonly used "M of N" method. The new method provides substantial improvement in the false track initialization rate and the average latency for initializing target tracks.