Impact of Imperfect Source-to-Relay CSI in Amplify-and-Forward Relay Networks

Cooperative communication assisted by relays is regarded as an effective means of achieving spatial diversity. The achievable diversity gain relies on the channel-state information (CSI) that can be acquired at the destination because spatial diversity combining unavoidably requires the disintegrated channel information. To reduce signaling overhead, quantization and compression are applied prior to the delivery of the source-to-relay (SR) CSI to the destination. This paper studies an amplify-and-forward relay network while considering the impact of the quantization of the SR CSI. Several quantization methods relying on the statistics of the SR CSI are investigated. The probability density functions of upper and lower bounds on the signal-to-noise ratio (SNR) and the corresponding symbol-error rate (SER) achieved at the destination are derived analytically. Monte Carlo simulations are performed to compare the simulation results with the analytical results of the SER bounds and to verify the effectiveness of the proposed channel quantizer. A combined method of determining the quantization interval based on equal-probability partitioning and determining the representative quanta based on the minimum-mean-square-error criterion is proposed and is verified to have the lowest quantization error among the noniterative quantization methods compared here. It is shown via the simulations that when using an iterative procedure based on the Lloyd-Max procedure, three bits per dimension are sufficient to achieve negligible SNR degradation.