Information-Aware Joint Calibration of Microphone Array and Sound Source Localization

Accurate calibration of microphone arrays is essential for various applications. However, existing graph SLAM-based methods, which introduce additional poselandmark constraints to enhance calibration accuracy, face challenges related to data redundancy and increased computational costs. In this paper, we propose a novel approach that combines the graph SLAM framework with an information-theoretic data selection strategy for efficient joint microphone array calibration and sound source localization. Our method evaluates and selects the most informative data segments by leveraging mutual information, effectively filtering out redundant measurements. This approach results in a streamlined and efficient calibration dataset that facilitates high estimation accuracy while significantly reducing the computational burden. Extensive simulations and real-world experiments validate that our method outperforms existing full batch data processing methods, demonstrating its potential for realtime applications. All the codes and datasets are publicly available at https://github.com/AISLAB-sustech/IA-Microphone-Calibration.