Automaatne Mobiilseadmete Ühendumine Integreeritud Mikrofonide Abil
Laen...
Failid
Kuupäev
Autorid
Ajakirja pealkiri
Ajakirja ISSN
Köite pealkiri
Kirjastaja
Abstrakt
Käesolev bakalaureusetöö kirjeldab meetodit, kuidas mobiilseadmete sensorite abil saada informatsiooni kasutajate sotsiaalse konteksti kohta. Seadmete mikrofone kasutades tehakse kindlaks, millised kasutajad paiknevad samas asukohas. Selle saavutamiseks viiakse sensorandmete peal läbi klasteranalüüs välisel serveril. Lisaks analüüsile annab server ka koos asuvatele seadmetele teada üksteise olemasolust ning varustab neid juhistega, kuidas teineteisega Bluetooth-i abil ühenduda. Käesolev töö pakub sellise platvormi teostuse kirjeldust. Testitulemused näitavad, et pakutaval lahendusel on koos asumise kindlaks tegemisel 75% täpsus.
This thesis proposes a method for using integrated mobile device sensor data to gain information about users' social context. In-device microphones are utilized to determine which users are co-located (for example, participating in a business meeting). This is achieved by clustering the sensor data at a remote server which also notifies the co-located devices of each others' existance and provides instructions on how to communicate with one another using Bluetooth technology. The work contributes a description of an entire implementation for such a framework. Test results show an accuracy of 75% in recognizing user co-location.
This thesis proposes a method for using integrated mobile device sensor data to gain information about users' social context. In-device microphones are utilized to determine which users are co-located (for example, participating in a business meeting). This is achieved by clustering the sensor data at a remote server which also notifies the co-located devices of each others' existance and provides instructions on how to communicate with one another using Bluetooth technology. The work contributes a description of an entire implementation for such a framework. Test results show an accuracy of 75% in recognizing user co-location.