Balso kokybė mobiliojo ryšio sistemose

Abstract

Šiame baigiamajame magistro darbe tiriama galimybė prognozuoti balso kokybės įvertį naudojant radijo ryšio kanalo parametrus. Prognozei panaudoti du metodai: tiesinė prognozė ir dirbtiniai neuroniniai tinklai. Tyrimas atliktas Vilniuje, 2G tinkle. Išanalizuoti matavimų metu gauti rezultatai, padarytos išvados. Prognozė atlikta remiantis trimis radijo kanalo parametrais: klaidingų bitų įvertis BER (angl. Bit Error Ratio), nešlio signalo lygis virš interferencijos C/I (angl. Carrier / Interference), signalo lygis RSSI (angl. Receive Signal Strength Indicator). Darbe patobulinta studento sukurta balso kokybės matavimo įranga. Gauti rezultatai palyginti su standartizuota TEMS Investigation 9.0 įranga, įvertintas sukurtos įrangos tinkamumas.

This master thesis was carried out during the study, which investigated the possibility of predicting voice quality estimate using radio channel parameters. Predictions using two methods: linear prediction and artificial neural networks. The study was conducted in Vilnius, the 2G network. To analyze the measurement results obtained, and conclusions. Prediction made on the basis of three radio channel parameters: bit error estimate BER (Bit Error Ratio), the carrier signal level above the interference C/I (Carrier / Interference), the signal level of RSSI (Receive Signal Strength Indicator). Improved student work created in the voice quality measurement system. The obtained results were compared with standardized TEMS Investigation 9.0 software. The conclusions drawn about the suitability of equipment designed for.