Abstract :
In this document, our latest work on distance estimation algorithm of acoustic based shot estimation system will be discussed. Information about our 4 point sensor array and the omni-directional microphones will be provided. Using Gaussian Mixture Models, a voice activity detection front end is designed. When a shot is detected, critical raw data is processed for directional measurement using generalized cross correlation based time difference of arrival. Applying directional calculation for both the muzzle blast and the shockwave, shot distance is estimated. Under the assumption that the bullet has speed and ideal environmental conditions, according to MATLAB simulations, depending on the miss distance, on average 5,6% distance calculation error is detected for distances up to 1000 meters. For limited amount of real audio data, the usability of this approach is discussed for 300m shot distance and a 10m miss distance.
Keywords :
Gaussian processes; acoustic measurement; acoustic signal processing; distance measurement; microphones; mixture models; sensor arrays; time-of-arrival estimation; Gaussian mixture models; MATLAB; acoustic based shot estimation system; cross correlation based time difference of arrival; directional calculation; directional measurement; distance 10 m; distance 300 m; distance calculation; distance estimation algorithm; omnidirectional microphones; sensor array; shot estimation systems; voice activity detection front end; Acoustics; Conferences; Estimation; MATLAB; Mathematical model; Reactive power; Signal processing; acoustic source localization; shot distance detection; time difference of arrival;