Multipurpose microphone array using Raspberry Pi and MATRIX Creator

Abstract

Microphone array have seen widespread usage over the years due to their effectiveness in fields of speech enhancement and speech recognition. The popularity of Smart Home Assistant devices, such as Google Nest, equipped with a microphone array, has proved the usefulness of a microphone array systems in speech related applications.

The lack of customizable features, inflexibility in deployment and high cost of entry present in commercial microphone array systems nowadays impedes the research and development into microphone array systems.

This project proposes a solution for the development of a microphone array system, utilizing the MATRIX Creator and Raspberry PI as the primary hardware components, due to the customizability, flexibility and portability that they provide. A cross mobile application was developed to serve as the central control system, to provide hands-free control over the hardware component of system. The microphone array system follows a wireless architecture, allowing the mobile application and multiple microphone array devices to be interconnected for synchronous control. This project shall also include the development of a severless application for the purpose of applying speech related techniques on recorded audio files.

The functionalities defined in this project are developed to be capable for application in a real world environment. With the heavy computational components required by the application of speech related techniques being isolated from the central control system of the microphone array system, computational and memory overheads are kept to the minimum.

Despite the challenges faced, this project has achieved its objectives of developing a microphone array system that supports the ease customization and deployment in real world environments while ensuring that there are minimal computational overheads across components.