A fast 3D-MUSIC method for near-field sound source localization based on the bat algorithm

Abstract

To improve the computation and real-time performances of the multiple signal classification (MUSIC) algorithm in 3D space, a fast sound source localization method based on the bat algorithm (BA) and the 3D-MUSIC, called BA-based 3D-MUSIC algorithm (3D-BMUSIC), is presented in this paper. 3D-BMUSIC greatly reduces the computation load by replacing the regular grid search with the BA. First, the near-field spherical wave model is established to obtain the spectral function of the 3D-MUSIC. Then, the spectral function is defined as the fitness function, which calculates the fitness value corresponding to each bat position. Finally, the global optimal bat position with the largest fitness value, as sound source localization, is obtained by successive iteration and sorting. The simulation and experiment show that 3D-BMUSIC accurately estimates the DOA and distance of near-field sources, and the root-mean-square error (RMSE) of 3D-BMUSIC is less than that of 3D-MUSIC. In addition, 3D-BMUSIC effectively reduces the computation time by approximately 96–98%. With shorter computation time and higher efficiency, 3D-BMUSIC promotes hardware implementation and is more suitable for high-precision localization of near-field sound sources.

Multisound source joint localization algorithm for moving targets with multipath coherent sources based on DOA clustering algorithm

“…Literature [3] uses the global optimization search ability of bat algorithm to analyze the characteristics of multi-path coherent source based on the joint localization algorithm of multi-path coherent source moving target and multiple sound sources based on bat algorithm, and finds out the optimal location of sound source by simulating the variable amplitude echo positioning behavior, so as to improve the positioning accuracy. However, this method depends on the distribution of the initial population.…”