Acoustic Interference System for Audio Privacy Protection

The proliferation of smart IoT devices with always-on microphones creates significant privacy vulnerabilities through unauthorized audio recording. This project presents an "Acoustic Interference System for Audio Privacy Protection" that prevents eavesdropping by leveraging the inherent hardware nonlinearity of commercial off-the-shelf (COTS) microphones. The system utilizes an array of ultrasonic transducers to emit near-ultrasonic waves (20–40 kHz) that are inaudible to human ears. Upon reaching a target microphone, these signals undergo self-demodulation at the diaphragm, generating a low-frequency "shadow signal" within the audible spectrum. This induced noise effectively masks legitimate speech, rendering recordings unintelligible to human eavesdroppers and automatic speech recognition engines. By targeting the physical layer of the recording hardware, the system provides a robust defense that cannot be easily bypassed by software filters. Experimental evaluations demonstrate that the system successfully disrupts unauthorized recordings while maintaining a quiet user environment Keywords — IoT, Eavesdropping, nonlinearity, Ultrasonic Transducers, Self-demodulation, Shadow signal.