Active user detection and channel estimation in uplink CRAN systems

Cloud Radio Access Network (CRAN) is proposed as a promising network architecture for future mobile communications. In this paper, we consider the topic of active user detection (AUD) and channel estimation (CE) in uplink CRAN systems with sparse active users. Different from conventional AUD and CE approaches which require the length of uplink pilots to scale with the number of users times the number of antennas per user, a novel algorithm will be proposed to substantially reduce the uplink training overhead by leveraging the technique of compressive sensing (CS). To achieve this goal, we first transform the problem of AUD and CE into standard CS problems. We then propose a modified Bayesian compressive sensing (BCS) algorithm to conduct AUD and CE in CRAN, which exploits not only the active user sparsity, but also the innate heterogeneous path loss effects and the joint sparsity structures in multi-antenna uplink CRAN systems.