Dynamic spectrum refarming for GERAN/EUTRAN considering GERAN voice traffic

This research article has been conducted for the efficient use of frequency spectrum in multi-RAT cellular network using dynamic spectrum refarming (DSR). For this purpose, two overlapping networks, GERAN and EUTRAN, and their voice and data traffic statistics are taken into consideration. Our investigation of hourly traffic changes in the voice service in the GERAN and the data service in the EUTRAN (which are the dominant services of each of these networks), indicates that these changes are approximately orthogonal. This issue is the main motivation for use of DSR methods, including DSR without inband/guardband (I/G) overlay and DSR with I/G overlay, for efficient assignment of frequency resources to GERAN and EUTRAN. In this paper, the main focus is on the process of performing DSR with I/G overlay in a real network and its challenges. For this purpose, we used a nonlinear autoregressive neural network to predict the traffic changes of voice service in the GERAN. By this method, the prediction error is less than 6.5% in the peak hours. Also, a method for efficient use of GERAN and EUTRAN carrier numbers (ARFCN and EARFCN, respectively) has been proposed to reduce the mutual interference in DSR with I/G overlay. The results show the significant impact of the DSR with I/G overlay on increasing the EUTRAN average bit rate compared to the classic spectrum refarming method (168% in downlink and 146% in uplink).