Opportunistic feedback and scheduling to reduce packet delays in heterogeneous wireless systems

We consider an opportunistic feedback and scheduling scheme for multiuser wireless systems. In opportunistic feedback schemes, a set of thresholds is assigned to the users, and only users with channel quality above the threshold transmit the feedback on channel state information (CSI) to the base station. We propose an opportunistic feedback scheme that combines techniques from queue/weight-based opportunistic scheduling to improve users' delay performance. A set of queue-based weights designed to slowly track variations in users' backlog is computed; then, each user is assigned a threshold based on the weights. The proposed scheme effectively captures heterogeneity in the offered loads, which is crucial for delay performance. To that end, we formulate a nonconcave maximization problem and propose an approximation algorithm that is numerically shown to have high accuracy. By simulation, we show that our scheme substantially reduces the mean delay, compared with conventional schemes.