Gunshot detection system for JTRS radios

Abstract

This paper presents a novel solution for detecting and locating gunshots using minimal equipment on a JTRS radio is proposed. Each radio acts as a sensor node to determine and share muzzle blast time of arrival information in order to determine a shooter location. The feasibility of such a system was investigated by developing detection and location algorithms and testing each algorithm's effectiveness in both simulation and with real data. The detection algorithm is used to accurately pinpoint the arrival of the muzzle blast of a gunshot at a single microphone and was realized using a rake-correlation filter loop. The location algorithm used the position and time of arrival information gathered from multiple sources to determine the shooter location and was realized by an extended time-invariant Kalman filter. The Kalman filter is very adaptable and can incorporate additional measurements or sensors such as terrain information to eliminate large vertical errors; or Angle of Arrival information to incorporate additional gunshot systems into the network. Results showed that in realistic scenarios, with only time of arrival information, gunshot location accuracy is dictated by radio positioning.