Biologically Inspired Acoustic Source Localization System

Situational awareness is a critical factor for the soldier in the field. An acoustic source localization system can greatly benefit the soldier by detecting threat and displaying its location. The goal of this proposal is to study a promising algorithm based on a biologically inspired technology and verify its ability to perform in various environments and with multiple sources. The proposed source localization technology involves a two-stage localization procedure to detect, classify and localize sources. A Constant False Alarm Rate sound detection algorithm is used to detect the presence of an acoustic event, classify it as noise or signal and determine its coarse location. A second-stage time-delay algorithm is used to locate sound sources with greater precision. Results of the localization algorithm will be presented using 3D auditory displays and a visual representation. The full system will be evaluated in a multi-source dynamic environment with selected acoustic source material that spans a representative range of potential sounds under a variety of noise conditions. The expected result is a functional system that will automatically detect and localize acoustic source material as well as, or better than, the human auditory system.