Distributed Fronthaul Compression Design for Rateless Coded Multi-User Uplink Transmission in C-RAN

We consider the multi-user uplink transmission in cloud radio access network (C-RAN), which consists of two users, two radio remote heads (RRHs), and the baseband processing unit (BBU) pool. The users adopt Raptor code and send the modulated signals to both RRHs continuously until receiving the acknowledge character (ACK) feedback from the BBU. A novel fronthaul compression scheme for RRH is proposed. Explicitly, both RRHs quantize the received signal to meet the fronthaul capacity constraint. One of the RRHs further compresses the quantized signal with low density parity check (LDPC) code. Then not only the resulted syndrome bits, additional quantization bits are also sent to the BBU according to the remaining fronthaul capacity, which can reduce the quantization noise at the BBU. We optimize the LDPC code based on extrinsic information transfer (EXIT) analysis. Simulation results show that optimized LDPC code has good compression performance compared with regular LDPC code, and the proposed fronthaul compression scheme outperforms the conventional scheme without DSC.