Real-Time Wireless Sound Localisation and Frequency Analysis: A Low-Latency IoT Approach

This paper presents the design and implementation of a Cost-effective, directional audio monitoring system utilizing the ESP32 platform. To address the challenge of visualizing abstract acoustic phenomena, the system integrates a four-channel orthogonal microphone array with real-time Fast Fourier Transform (FFT) signal processing. The primary novelty of this work lies in the strategic synergy of the ESP32's high-speed dual-core architecture with the ESP-NOW connectionless protocol. This combination effectively overcomes the processing speed, memory, and wireless latency limitations found in alternative low-cost microcontrollers, such as standard Arduino or STM32 boards. By performing simultaneous 512-point FFT analysis on four independent inputs, the system achieves a frequency resolution of 15.625 Hz with a full-system update cycle of approximately 300ms. Experimental validation using tuning forks confirmed the system's accuracy in identifying dominant frequencies and source direction, attributing observed deviations to the FFT algorithm's inherent discrete "binning" rather than sensor error. Additionally, the system demonstrates sub-10 ms wireless transmission latency and identifies a positive correlation between sound frequency and reliable detection distance. This work establishes a robust, low-power foundation for intelligent acoustic sensing in IoT applications without the need for expensive digital signal processors.