Microphone Array and Raspberry Pi Interfacing for Real-time DOA Estimation and Tracking of Audio Sources

Abstract

Microphone arrays are commonly used for Direction of Arrival (DOA) estimation of audio sources. This paper presents the design and implementation of a real-time embedded system (RTES) for audio source localization. A commercially available microphone array assembly (consisting of four microphones) has been interfaced to Raspberry Pi 3B for estimating the arrival angles of audio/sound/acoustic sources. The acoustic signals received a teach of the microphones are used for DOA estimation using the Generalized Cross Correlation Phase Transform (GCC-PHAT) algorithm. The estimated DOA is then displayed on the serial monitor. Servo motors are interfaced to the Raspberry Pi and are programmed to point towards the estimated direction of the audio source (on the basis of the estimated DOA value). The estimated DOA value is also updated to the cloud-based server Firebase using the inbuilt Wi-Fi on Raspberry Pi, and can be remotely accessed through a smartphone running on android platform. The proposed system has been designed and tested successfully and it is observed that it provides accurate DOA estimation and audio source tracking in real-time. This has various uses in applications like voice assistant robots, machine audition, hearing aids, source remixing, augmented reality headsets, security cameras, sound tracking (sonar sensor) devices, bird identification, mammal localization and so on.