Angle of Accuracy Enhancement through FPGA-based Sound Source Localization with Multiple Sound Sensors

Sound Source Localization (SSL) refers to the process of determining the location of the sound. However, traditional SSL methods often suffer from limited accuracy, particularly in complex environments. This paper proposes an FPGA-based SSL system employing multiple sound sensors to achieve an enhanced angle of sound source arrival estimation. This system uses the Time Difference of Arrival (TDOA) to compare the time difference received by multiple sensors concerning the distance of sound sensors. In this work, the FPGA hardware acceleration enables real-time processing of sensor data, leading to improved accuracy through the design of SSL module, CORDIC module and CORDIC mapping module. A few experiments were conducted using a Micro-Electro-Mechanical System (MEMS) and Electret Condenser Microphones (ECMs) sound sensors. The experimental comparison shows that the MEMS microphone gives a more accurate angle of the sound sources with a 7° error of angle compared with the ECMs sound sensor that achieved a 27° error of angle. The percentage difference in angle inaccuracy between two and three sound sensors has increased by 38.72%.