Long-term window scheduling in multiuser OFDM systems based on large scale fading maps

In this paper resource management for non-realtime (NRT) services in multiuser orthogonal frequency division multiplexing (OFDM) systems with proportional fairness is investigated. Compared with previous algorithms, we allocate the subcarriers and transmit power based on the near-future channel conditions in a more cognitive manner. With this purpose we introduce the concept of a large scale fading (LSF) map, which is a geographical map with average LSF values. Combined with a mobility prediction model, the future channel conditions can be predicted. We present a long-term window scheduling algorithm, which iteratively allocates resources to all the users inside the shifting time windows in the future. We show that the proposed scheduling algorithm results in a significant improvement in the system sum rate compared to the previous scheduling algorithms. It´s also discussed that the accuracy of the positioning technique and the length of the scheduling window have an impact on the performance.