To reverse engineer the brain in order to discover novel design principles for physical systems providing enhanced information-processing capabilities for Navy/DoD and homeland security needs. Specifically, we are studying experimentally, and modeling mathematically, fundamental signal processing principles of neurons and neural systems for the purposes of designing and implementing acoustic pattern recognition systems for security breaching noise detection. Examples of security breaching noises of interest include: gunshots, RPG/mortar firings, sniper detection (chambering-a-round prior to firing), diesel truck engine sound, human voices (speaker-independent; language-independent), scaling chain-link fences, etc.
Benefits: Effective Situation Awareness requires quick, accurate, and effective information on events as they are unfolding. In the context of soldier safety in battle or convoys en route, the identification of pre-attack preparations, threats, or the location of source of attacks can make the difference in determining whether to take evasive actions or to engage the enemy. And those kinds of decisions have life effecting consequences. Non-line of sight acoustic smart sensors deployed in a network arrangement around areas needed to be secured, such as a convoy route, can play a vital role in recognizing telltale sounds that are either a precursor to an attack or the location of an actual attack weapon, such as an RPG or small arms.
Keywords: Smart Sensors, Acoustic Recognition, Threat Detection, Remote Sensors
