Acoustic Localisation of Gunshots in the Presence of Obstacles

Abstract

Acoustic localization of gunshots is of interest for protecting security camps in contested forests, increasing soldiers’ situation awareness in ambushes, as well as in guarding cities. All these scenarios are also fraught with obstacles coming in the path of acoustic waves. We present a computational study of the effect of obstacles on the accuracy of localization of point sound sources like gunshots. The two-dimensional linearized Euler equations are solved with the dispersion relation preserving scheme. The localization is performed using the wavefront curvature method based on pressure signatures recorded at three virtual microphones. We report at most 8◦ of error in the estimated bearing angle of the source across a range of practical values of five parameters – viz. the location and size of the obstacle, and the size, orientation and location of the sensor array relative to the source. This level of error may be quite acceptable in the scenarios considered. On the other hand, the range error is much more severe, exceeding 100% of the true range in a few cases. In the scenarios considered, the bearing angle estimate is much more critical than the range estimate. Overall, we conclude that acoustic localization approaches are quite robust in the presence of obstacles.