In English

Ultrasonic Source Localization with a Beamformer Implemented on an FPGA Using a High-density Microphone Array

Daniel Antonsson ; Mengxuan Li
Göteborg : Chalmers tekniska högskola, 2018. 57 s.
[Examensarbete på avancerad nivå]

A system has been designed which uses a high-density MEMS microphone array and a beamforming algorithm implemented on an FPGA to perform real-time positioning of ultrasonic sources. The array consists of 48 microphones which is a limit derived from the latency of the system; if the latency were to increase further the system could no longer be said to operate in real time.

The system calculates signal energy in front of the array and displays the energy on a heatmap which updates at a frequency of 5.2 Hz. It can locate near-monochromatic sources at 20 kHz and narrowband sources at 35 kHz. However, the ability of the system to locate these sources was proven to be unreliable due to a phenomenon called spatial aliasing which causes an ambiguous pattern in the heatmap. Very brief sounds could not be located by the system and the system is also easily disturbed by ambient noise, such as noise from conversations.

It was concluded that the first step of any potential future development should be to completely revise the implementation of the beamforming algorithm in order to increase the amount of microphones that can be used and decrease the latency of the system.

Nyckelord: microphone array, ultrasonic, FPGA, beamforming, delay-and-sum

Publikationen registrerades 2018-06-28. Den ändrades senast 2018-06-28

CPL ID: 255327

Detta är en tjänst från Chalmers bibliotek