Exploiting channel statistics to improve the average sum rate in OFDMA systems

The achievable average sum rates of both OFDMA uplink and downlink systems with statistical multiuser adaptation are investigated. We show that simultaneous utilization of channel mean and tap correlation information in allocating subcarriers and power to multilasers offers potentially large spectral efficiency gains compared to both uniform subcarrier and power allocation and multiuser adaptation exploiting channel mean information only. Such gains arise from the inherent multiuser diversity that the statistical adaptation can efficiently capture. The impacts of the Ricean K-factor, the magnitude of tap correlation, and the number of simultaneously scheduled users per OFDM symbol on the average sum rate are further investigated.