Use of Adaptive, Non-line of Sight Smart Sensors to Recognize and Locate Threats to Physical Assets and Field Forces

Biomimetic Systems will develop and test a simulated system to localize and identify target sound sources including pistol, shotgun, and rifle fire. The software simulation will be a behavioral model of a proposed hardware system for Phase II development. Our system extracts novel feature sets generated by a biomimetic processor based on human hearing, on which we train and evaluate a proprietary classifier algorithm. We propose to apply this approach to acoustic data collected in Phase 1. Field data will be collected using a vehicle mounted acoustic sensor array with live gunfire as well as samples of urban ambient noise, wind noise, and vehicle generated noise. Urban weapon fire scenarios with stationary and moving vehicles will be generated synthetically using collected field data to create test signals for the detection and localization algorithms. We will analyze the simulation results to determine system efficacy and identify possible failure modes. We anticipate based on prior work to localize sounds within 10 ms with accuracy on the order of 2 degrees in azimuth and 5 degrees in elevation. BENEFITS: The proposed technology will form the basis for a family of commercial products aimed at the civilian law enforcement and security markets. Acoustic sensing systems could be used by first responders to detect, classify and locate sources of gunfire in urban locations. Vulnerable facilities, like oil terminals, power plants, pipelines, powerlines, communications centers and the like could be continuously monitored for sound patterns linked to suspicious activities. Non-security applications are also envisioned in the automotive, aerospace, forestry and consumer markets.