This work analyses communication channel settings of UMTS technology, which are related with a data transfer throughput. The paper describes the measurement equipment that is suitable for measuring parameters of a mobile network channel. Besides, it analyses the suitability of this equipment and issue of parameter values that are associated with data throughput. Further, the study includes the selection of the most specific parameters, which are crucial for data speed. Using these parameters, models were developed for prediction of data transfer throughput. To build the model, the linear and nonlinear forecasting methods were used. The linear prediction was made by using linear regression, nonlinear — neural networks. Using the linear prediction model, 77.83% forecast accuracy was achieved, while the accuracy of forecasted nonlinear transmission rate amounted to 76.19%. The accuracy of prediction models was obtained using eight parameters of the communication channel. Finally, the paper presents the data throughput prediction models that allow predicting data speed with the help of channel parameters presented by a standard terminal. The list of these channel parameters is derived from five UE‘s of different manufacturers. The expected most accurate data transfer rate can be predicted using a set of parameters issued by Nokia device.

Article in Lithuanian