Fingerprint-based Sound Source Localization Using Iterative Interpolation Method

Abstract

The fingerprint-based sound source localization (SSL) approach usually requires tremendous time and efforts for location fingerprints collection in sampling phase and reference points (RPs) matching in positioning phase. In this paper, we propose an iterative interpolation method based on cluster analysis to reduce such calibration efforts and matching computation works. Unlike conventional interpolation methods, the novel method can make further efforts in refining the interpolation scope and monitoring the interpolation process to reduce the required virtual RPs. The proposed method significantly outperforms the conventional interpolation methods in efficiency on the premise of achieving the same or similar accuracy.