Journal of Modeling in Engineering

Journal of Modeling in Engineering

Sound source localization system using a microphone array, based on combining weighted features in a CNN-GRU model

Document Type : Research Paper

Authors
Department of Electrical Engineering, Faculty of Engineering and Technology, Lorestan University, Khorramabad, Iran
Abstract
Sound source localization (SSL) in noisy and reverberant environments remains a significant challenge, prompting extensive research into robust solutions. Recent advances in deep learning have demonstrated considerable potential for improving SSL performance. This paper presents a novel model that combines a convolutional neural network (CNN) with a gated recurrent unit (GRU) and a customized weighted feature fusion mechanism to enhance direction-of-arrival (DOA) estimation in acoustically complex conditions. The proposed CNN-GRU architecture extracts discriminative spatial features, including interaural time difference (ITD), interaural level difference (ILD), generalized cross-correlation phase transform (GCC-PHAT), and sound intensity from multichannel audio signals. To assess robustness and generalizability, two experiments are conducted, evaluating the model’s performance under varying additive noise levels and reverberation durations. Experimental results demonstrate that the proposed approach outperforms existing baseline methods, achieving a peak DOA estimation accuracy of 91.85% and a mean absolute error (MAE) of 2.46°, highlighting its effectiveness in challenging acoustic scenarios.
Keywords
Subjects


Articles in Press, Accepted Manuscript
Available Online from 27 October 2026

  • Receive Date 08 August 2025
  • Revise Date 17 April 2026
  • Accept Date 17 October 2026