Resource allocation algorithm in downlink of OFDMA systems based on users' classification and superposition coding

In this paper, we propose a simple proportional fairness resource allocation technique for downlink OFDMA system, by re-arranging the set of users into four classes and setting different priority order. The classes are obtained according to buffers states and channels conditions. We offer priority to users having a delay sensitive traffic by considering those who have the fullest buffer and good channel conditions. Moreover, we propose to use the superposition coding technique to increase the system capacity. In fact, we will allow a degraded user (having bad channel condition) and a potential user (having good channel condition) to share the same subcarrier. At the degraded user side, the receiver will proceed as usual and will consider the information sent to the potential user as an additive noise. At the potential user side, a successive decoding process will be adopted. To avoid error propagation, which can occur in this case, the amount of transmitted power to each user will be optimized accordingly. A suboptimal resource allocation algorithm, based on the mentioned superposition scheme and users classification, is derived. Simulation results confirm that the proposed technique offers a proportional fairness allocation and outperforms the classical algorithms in terms of total throughput.