Adaptive Processor for Large Area Urban-Noise Characterization, Signature Extraction, Detection, and Localization

Robust signature detection in noisy, inhomogeneous environments is depends upon extracting identifiable components from available data. Urban areas present a demanding setting; multi-path reflections, frequency-dependent attenuation, and local resonances degrade and mask the target signal. SR2 Group’s solution is based on proprietary processing that can improve signature detection and localization in target environments. Our technology is robust to distortion and coherence issues that confound conventional methods. It incorporates mechanisms that significantly reduce dimensionality without sacrificing decisive signal information. We attack four goals: extracting signature data from collections of known signals, dimensionality reduction, detecting signatures in novel signals, and localizing sources in space with ad-hoc arrays; we propose to demonstrate each. Phase I establishes viability of the processing methods and benchmarks performance in comparison to other classical and emerging methods, leading to recommendations for a Phase II demonstrator. SR2 Group has teaming relationships with The Norbert Wiener Center for Harmonic Analysis and Applications at the University of Maryland and with a national electronics device manufacturer. The effort will achieve a processing architecture tailored to wide area urban surveillance. Proposed technology is also applicable to unattended ground sensors, event-alert detectors, and other acoustic monitoring systems where robust, autonomous signature detection is demanded.