guest ::
| Home > Publications database > Closed-loop sound source localization in neuromorphic systems |
| Home > Publications database > Closed-loop sound source localization in neuromorphic systems |
| Journal Article | FZJ-2024-01287 |
; ; ; ; ; ;
2023
IOP Publishing Ltd.
Bristol
This record in other databases: 
Please use a persistent id in citations: doi:10.1088/2634-4386/acdaba doi:10.34734/FZJ-2024-01287
Abstract: Sound source localization (SSL) is used in various applications such as industrial noise-control,speech detection in mobile phones, speech enhancement in hearing aids and many more. Newestvideo conferencing setups use SSL. The position of a speaker is detected from the difference in theaudio waves received by a microphone array. After detection the camera focuses onto the locationof the speaker. The human brain is also able to detect the location of a speaker from auditorysignals. It uses, among other cues, the difference in amplitude and arrival time of the sound wave atthe two ears, called interaural level and time difference. However, the substrate and computationalprimitives of our brain are different from classical digital computing. Due to its low powerconsumption of around 20 W and its performance in real time the human brain has become agreat source of inspiration for emerging technologies. One of these technologies is neuromorphichardware which implements the fundamental principles of brain computing identified until todayusing complementary metal-oxide-semiconductor technologies and new devices. In this work wepropose the first neuromorphic closed-loop robotic system that uses the interaural time differencefor SSL in real time. Our system can successfully locate sound sources such as human speech. In aclosed-loop experiment, the robotic platform turned immediately into the direction of the soundsource with a turning velocity linearly proportional to the angle difference between sound sourceand binaural microphones. After this initial turn, the robotic platform remains at the direction ofthe sound source. Even though the system only uses very few resources of the available hardware,consumes around 1 W, and was only tuned by hand, meaning it does not contain any learning atall, it already reaches performances comparable to other neuromorphic approaches. The SSLsystem presented in this article brings us one step closer towards neuromorphic event-basedsystems for robotics and embodied computing.
; 
; 
; Article Processing Charges ; DOAJ Seal ; Fees
|
The record appears in these collections: |
PDF