Blind Detection, Separation, and Location of Dense Co-Channel Emitters using Multiplatform Spatial-Coherence Restoral

The multiplatform spatial-coherence restoral (MP-SCORE) algorithm, introduced in 1988 to blindly separate spatially-coherent, temporally featureless signals based on their differing TDOA and FDOA at spatially separated antenna arrays, is extended to blind detection, separation, and location of dense co-channel emitters using wide-baseline collection networks. The original MP-SCORE algorithm is reviewed and related to the ML estimator of the location of a Gaussian-distributed source (or of the location and content of an unknown source) received in the presence of independent Gaussian-distributed interference with arbitrary spatial covariance at each array in the network. False-alarm and miss-rates for the corresponding MP-SCORE detector are also derived, and are shown to be a function of only the time-bandwidth product of the detector and the maximum source SENR attainable at each array in the network. Practical methods for detecting, separating, and locating multiple co-channel sources using MP-SCORE are then presented, and the end-to-end algorithm is demonstrated for a GPS C/A code jamming scenario in which two 8-sensor airborne receivers attempt to detect and locate 100 narrowband (2 MHz) noise-loaded emitters continuously transmitting over a wide (150times150 nmi) geographical region at the center of the GPS L1 frequency band. The simulated algorithm detects and geolocates 27 emitters to within 1/2 nmi over a single 26 ms snapshot, and detects and geolocates 38 emitters to within 445 feet over eleven 26 ms snapshots collected at 15 nmi intervals along the edge of the emission region