Improved Target Discrimination of Multiple Targets Using Bulk Filtering for Debris

Wavelet Technologies, Inc. (WTI) proposes development of algorithms that process radar returns from dense threat complexes and attempt to discriminate between actual threats, countermeasures, and incidental debris generated from the rocket motor burnout and associated events. We term these objects collectively a Low Relative Velocity (LRV) debris field, in which the threat is embedded. Because practical radars cannot resolve objects in the LRV debris field, algorithms must operate only on range and Doppler data for determining the radar track. We propose to use expectation maximization (EM) algorithm that work on this basis to derive suitable object tracks without rejecting tracks for threats because a strict rejection threshold for tracks is not employed. These tracks will then be processed using a Bayesian classifier or a kernel Bayesian classifier to discriminate between threats and other objects in the LRV debris field. Simulations of this process will be conducted in Phase I using NASA"s orbital debris evolution program on modeled LRV debris. Radar observations will be taken at several simulated sites and combined in the tracking program to address data acquisition from multiple radar platforms.