Locating Acoustic Sources with Multilateration: Applied to Stationary and Moving Sources
Translated title
Lokalisering af akustiske kilder: Anvendt på stationære og dynamiske kilder
Author
Dalskov, Daniel
Term
4. term
Education
Publication year
2014
Submitted on
2014-06-04
Pages
76
Abstract
Dette speciale undersøger, hvordan man kan finde, hvor en lyd kommer fra. Metoden bygger på multilateration, hvor tidsforskellen mellem ankomsten af lyden til flere mikrofoner (TDOA, Time Difference of Arrival) bruges til at beregne en position. Den kan lokalisere både stationære og bevægelige kilder, når der arbejdes i et anekoisk (ekko-frit) miljø. For at muliggøre online eksekvering er der udviklet en hardwareplatform baseret på en Raspberry Pi, et Wolfson Audio-kort og forforstærkere til mikrofonerne. Et kommunikationsproblem mellem Raspberry Pi’en og lydkortet begrænsede dog antallet af indgangskanaler til to i stedet for de tre, der kræves til 2D-lokalisering. Derfor kan metoden på denne platform kun bestemme lydkildens vinkel (retning) og ikke dens fulde position.
This thesis investigates how to determine where a sound comes from. The method uses multilateration, comparing when a sound reaches several microphones (time difference of arrival, TDOA) to compute a position. It can locate both stationary and moving sources when used in an anechoic (echo-free) environment. To enable online execution, a hardware platform was built using a Raspberry Pi, a Wolfson Audio Card, and microphone preamplifiers. However, a communication issue between the Raspberry Pi and the audio card limited the number of input channels to two instead of the three needed for 2D positioning. As a result, on this platform the method can only estimate the source’s angle (direction), not its full position.
[This abstract was generated with the help of AI]
Keywords
Documents