Artificial Neural Network Classification for Gunshot Detection and Localization System


Authors : Cherry Mae J. Galangque, Sherwin A. Guirnaldo.

Volume/Issue : Volume 3 - 2018, Issue 6 - June

Google Scholar : https://goo.gl/DF9R4u

Scribd : https://goo.gl/G9LMVj

Thomson Reuters ResearcherID : https://goo.gl/3bkzwv

At present, better situational awareness is undeniably paramount for a soldier in the field, whether he is dismounted or on board a vehicle. Quick determination of where the enemy fire comes from could save life in a very deadly situation. In military setting, there is a long-felt need for a system and a method to identify acoustic events such as gunshots and to locate the shooter’s location as quickly as possible. We envisioned a system to detect gunshots and to locate the shooter’s location. Also, friendly boots on the ground near the sensor carrier can receive enemy position data through WiFi or Bluetooth connectivity with their smart devices (e.g., smartphones). With this, soldiers will have the technology that will assist them in knowing what happened and where the shooting came from. The study employed the adaptive capability of an Artificial Neural network in the detection and localization process. Microphones were used as primary sensors. The implemented system was subdivided into two modules: the classification and localization module. Sound signal properties were used to identify or differentiate gunshot from background noise and other explosive acoustics; difference in time of arrival and signal strengths were used to locate the origin of the gunshot sound. Around thousand rounds of 5.56mm where used to qualify the performance of the system. To test the 360 degrees performance, the sensor set up was gradually rotated while the position of the shooter was incrementally increased by 10 meters starting from a distance of 50 meters up to 100 meters. A success rate of around 99 percent is guaranteed in distinguishing sound from M-16 rifle from that of background noise or fire crackers. On the other hand, test result for localization showed that the system is capable of providing more than 90 percent accuracy for the source orientation, i.e., assuming +/- 15 degrees precision. Distance accuracy of more than 90 percent was also observed during the test (assuming +/- 5m precision).

Keywords : Artificial neural network, classification, localization, acoustic, gunshot detection.

CALL FOR PAPERS


Paper Submission Last Date
30 - September - 2021

Paper Review Notification
In 1-2 Days

Paper Publishing
In 2-3 Days

Never miss an update from Papermashup

Get notified about the latest tutorials and downloads.

Subscribe by Email

Get alerts directly into your inbox after each post and stay updated.
Subscribe
OR

Subscribe by RSS

Add our RSS to your feedreader to get regular updates from us.
Subscribe