Training Design and Channel Estimation in Uplink Cloud Radio Access Networks

Abstract

To decrease the training overhead and improve the channel estimation accuracy in uplink cloud radio access networks (C-RANs), a superimposed-segment training design is proposed. The core idea of the proposal is that each mobile station superimposes a periodic training sequence on the data signal, and each remote radio heads prepends a separate pilot to the received signal before forwarding it to the centralized base band unit pool. Moreover, a complex-exponential basis-expansion-model based channel estimation algorithm to maximize a posteriori probability is developed, where the basis-expansion-model coefficients of access links (ALs) and the channel fading of wireless backhaul links are first obtained, after which the time-domain channel samples of ALs are restored in terms of maximizing the average effective signal-to-noise ratio (AESNR). Simulation results show that the proposed channel estimation algorithm can effectively decrease the estimation mean square error and increase the AESNR in C-RANs, thus significantly outperforming the existing solutions.