FPGA implementation of Acoustic Source Localization algorithm

Abstract

Acoustics Sensing and Intelligent Acoustics has produced revolutionary changes in the area of Audio Signal Processing. Acoustic Source Localization has a wide range of scope in applications such as Robotics, Security Surveillance, Speech quality Enhancement in hands free Telephony, Aeronautics to track Low flying aircrafts and Video Conferencing for Military congruence like SONAR. The Proposed system aims at obtaining the Directionality of Acoustic signal and estimating the Azimuth angle. All the functions for the Acoustic source localization are implemented with the use of a dedicated parallel architecture in FPGA. The sound signals captured by microphones at different positions are taken as input database. Those sound signals may be affected by any environmental disturbances or noise. So, the sound signals are preprocessed using Filter for Noise Cancellation purposes. The Preprocessed sound samples are fed as a input to the Sound Localization Module. Time Delay Of Arrival (TDOA) technique is employed to obtain the directionality of the Acoustic signal. The Time delays between the pairs of Sound signals are computed effectively using Generalized Cross Correlation Algorithm. Using the Time delay, the Position and the Azimuth angle are retrieved. The Localization algorithm is implemented in Virtex-5 FPGA as it promotes High Parallelism.