Semi-dynamic user-specific clustering for downlink cloud radio access network

This paper studies user-specific clustering for a downlink cloud radio access network (C-RAN), where a central unit (CU), connected to all base stations (BSS) via limited-capacity backhaul links, coordinates the BSS to form cooperative clusters for every user. By taking into account the training overhead for channel estimation in C-RAN, we design the clustering scheme aimed at maximizing the average net throughput of the network subject to the constraint on backhaul capacity, where a hybrid coordinated multipoint (CoMP) transmission mode is considered. The proposed clustering scheme can be operated in a semi-dynamic manner merely based on large-scale channel information, has low computational complexity, and performs close to the optimal scheme found by exhaustive searching. Under two special cases where the backhaul capacity is very stringent and unlimited, the proposed scheme is then tailored for pure coordinated beamforming (CB) mode and pure joint transmission (JT) mode to further reduce the clustering complexity. Simulation results show that the proposed semi-dynamic clustering schemes are superior to the dynamic clustering scheme due to the reduction of required training overhead.