Mennitt D., Gillett P., Carneal J., Johnson M.

The study of distributed sensor networks is a rapidly growing field with a wide array of applications. A network of mobile microphone arrays (nodes) has been created to identify and track noise sources in an outdoor environment. Collaborative signal processing is exploited: individual nodes with limited processing power and access to the local environment relay information to a central-level composite tracking system (fusion center). The nodes are able to detect noise sources, calculate the relative direction, and perform gross classification. The fusion center is responsible for associating detections to form noise events, associating events to objects, and then tracking those objects. Association is accomplished by a series of classification algorithms and decision logic. Information is stored on the objects’ location, speed, heading, track, and noise character: impulsive or tonal, amplitude, timbre, etc. Results from experiments run in an outdoor suburban environment demonstrate performance and the complexity of the problem. Challenges and directions for future research are discussed.