Phased array acoustic sensing for sound localization using beamforming and principal component analysis

Applications of acoustic sensing have become increasingly popular, particularly within multimodal sensing platforms. This increase in popularity is driven by the need for greater situational awareness, environmental monitoring, and improved detection of acoustic events in various settings. This paper presents an implementation and analysis of a uniform rectangular array (URA) for environmental monitoring using an FPGA. The project uses beamforming with a principal component analysis (PCA) detection scheme to create a long range monitoring system with a wide field of view. We discuss the design considerations of acoustic beamforming, detailing the FPGA architecture used for audio data collection and covers the processing techniques employed to filter audio from the phased array. The beamformed audio demonstrated a signal increase similar to simulations. The PCA detection scheme was able to successfully detect vehicles from up to 1km away at different angles. The results demonstrate the potential for enhanced acoustic sensing using beamforming with PCA.