A New Algorithm Of Sound Source Localization Based On Special Acoustic Detection

The technical research of sound source localization is a rising technology subject that involves many technology areas, such as Acoustics, Digital Signal Processing, and Intelligent Optimization Algorithms. This research plays an important application value in video conference and target location. Especially for counter-snipers, localizing the3-D position of the sound source efficiently is of great importance to control the situation and to safeguard lives.As a subtopic of a project that commissioned by a Nanjing communication technology co.,LTD, this subject is based on technology of signal acquisition, detection, and identification process. It aims to establish a portative microphone array model that helps to construct algorithm of sound source localization. Finally, this kind of algorithm precisely localizes strong sound source signal that detected and identified during earlier stage.This paper focus on researching gunfire source localization, giving preference to localization algorithm of the Time Delay of Arrival(TDOA) through studying the technology of sound source localization and combining the character of sound signal of gunshot source.This paper gives details to introduce algorithms of Least Mean Square Time Delay Estimation (LMSTED) and Generalized Cross Correlation Time Delay Estimation (GCCTED). By the numerical simulation of two kinds of algorithms, the result shows that GCCTED Algorithm has a high precision with fewer computing time. In order to utilize the TDE value of weight function efficiently, this paper proposes GCCTDE algorithm of Minimum-variance Unbiased Estimator. The simulation result proves precision have improved obviously.Besides, in terms of Direction of Arrival (DOA), this paper establishes a3-station microphone array model that differs from traditional models and DOA Estimation Algorithm. The simulated experiment shows that this algorithm has a high locating precision and adapts to remote environment.