Gunshot Classification and Localization System using Artificial Neural Network (ANN)

A number of reports about ambush among soldiers costing most of their lives have been recently reported as they try to uphold peace and safety among the individuals living in a certain place. Knowing where the gunshots came from is an advantage for the soldiers to have better aim. The purpose of the study is to aid soldiers in times of combat to avoid relatively large unpredictable loss and to safeguard our territories from unconscious attacks by developing a gunshot classification and localization system. This study utilized the adaptive ability of an Artificial Neural network (ANN) in the classification and localization of gunshots using microphones as the primary sensors. The system consists of two modules: the classification and localization module. Sound signal properties were used to distinguished gunshot from background noise and other explosive acoustics such as firecrackers; the difference in time of arrival and signal strengths of the muzzle blast and shockwave were used to locate the origin of the gunshot sound. Around a thousand rounds of M16 gunshots where used to qualify the performance of the system. The system achieved around a rate of 99 percent in distinguishing M16 gunshots from that of background noise or firecrackers. On the other hand, the test result for localization showed that the system is capable of providing more than 90 percent accuracy for the source angle orientation and more than 90 percent was also observed for the source distance orientation.