FPGA Implementation Of Sound Source Localization Algorithm Based On Microphone Array

Sound source localization is to find the spatial location of the speaker. It is one of the array signal processing key technologies based on microphone and has wide applied foreground.Sound source localization that is realized by dealing with sound signals picked up by microphone array, can be used to determine and track speaker. It is a new hotspot in acoustic signal processing field and a research having practical signification. With the development of the research, the technique has been widely used in fields such as video conference, telephone conference, intelligent robot, voice check and speech recognition etc. Yet, the algorithm realization based on FPGA is rarely seen. This paper describes and realizes it on FPGA. The following work has been done:First of all, the basic theories of main kinds of sound source localization technologies have been summarized and compared. Also the characteristics of speech signal process and the basic theory and method of acoustic signal pre-process are discussed. Then the basic principle of TDOA (time delay of arrival) is given and a geometries localization method is described in detail, which is applied to estimate the plane angle and the polar angle. The method based on TDOA is the most popular algorithm in sound source localization which consists of two steps, that is, time delay estimation and localization estimation.Then, according to the forgoing introduction and analysis, a realizable sound source localization method is proposed. Finally, sub modules are completed by hardwire description language (HDL). Some optimization methods are applied to them, and experimental result and timing simulation waveforms are listed in the end. The sub modules include: band-pass filter, adding window and frame partition, endpoint detection, time delay estimation, location estimation etc.