SBIR-STTR Award

SSPI proposes to develop a SIGINT tool that would produce high-accuracy emitter position estimates for a large class of measurement types including propagation time (PT), time difference of arrival (TDOA), frequency difference of arrival (FDOA), and angle of arrival (AOA). If the source is known to be near the surface of the earth and a digitized terrain map is provided, this information can be used to enhance the quality of the estimate. The robustness and general applicability of the proposed Maximum Likelihood approach make it ideally suited for high-accuracy geolocation through the combining of multiple geo-observables from distinct types of SIGINT sensors. The proposed general-purpose geolocation technique can be used for law enforcement and emergency services to locate a transmitter. The technique is also applicable for FCC to locate unauthorized transmitters. Another potentional application is to locate malfunctioning or modified radios in a cellular or PCS communications system.

The objectives of the R&D proposed are to develop and evaluate innovative methods for fusing data from multiple diverse distributed sensors to perform high-accuracy emitter geolocation in the presence of multipath propagation and possible cochannel interference. The proposed methods are optimum subject to constraints that enable programmable tradeoff of required bandwidth and covertness (of each sensor's data transmission) for emitter location accuracy. The approach assures efficient cost effective use of collection assets. Many of the algorithms comprising the proposed geolocation system have been specified, and some have already been evaluated. This project will result in software prototypes for the emitter geolocation system for battlefield applications, and for a CCI Analysis and Source Location subsystem for commercial use in cellular drive-test equipment for network monitoring, analysis, and optimization.

Keywords:
GEOLOCATION, MULTIPATH, INTERFERENCE, COVERTNESS, MULTISENSOR, DATA-FUSION, MODEL-FITTING, MAXIMUM-LIKELIHOOD