Efficient Channel Quality Feedback Schemes For OFDMA Systems With Different Schedulers

In orthogonal frequency division-multiple access (OFDMA) systems, the available downlink frequency is divided into subbands which are shared among users. The transmitter determines the users that are scheduled on each subband, and the data rate for each user. To enable these transmitter choices, each user feeds back a channel quality indicator (CQI), giving a compressed version of its observed signal-to-interference-noise ratio (SINR) on all subbands. In this paper, we classify existing CQI compression schemes into symmetric and asymmetric schemes. Symmetric schemes aim to accurately represent the CQI on all subbands, and are shown to be suitable when the receiver cannot accurately predict the subbands on which it will be scheduled. This happens, for instance, when the receiver is fast-moving, or when a round- robin scheduler is employed, or during low load conditions. Asymmetric schemes, on the other hand, greedily achieve accuracy only in high-SINR subbands, and are shown to perform well in high-load cellular networks with proportional fair scheduling and low-speed users. We propose one new scheme in each category. Simulation results for a 3GPP LTE cellular network show that our proposed schemes achieve similar or better performance than the current best-M feedback scheme used in LTE, with 30% less overhead.