Joint quantization strategy for uplink cooperative cellular interference alignment with limited feedback

We consider the uplink three-cell single-user cellular systems with limited feedback interference alignment (IA). Considering the exchange of channel state information (CSI) between base stations (BSs), IA solution can be done at BSs. Then the indices of the quantized precoding vectors are fed back to users. A joint quantization strategy, which quantizes the precoding vectors for different users jointly, is proposed. The proposed quantization scheme enlarges the codebooks and quantizes precoding vectors jointly through the exchange of codebook information between BSs. The quantized precoding vectors for different users are set to have a same index. Only one of the three BSs is appointed to broadcast the same index so that the total feedback cost would not be increased. With the proposed quantization scheme, the quantization errors are mitigated and the system sum rate is increased. A quantization criterion, which we refer to as the maximum sum rate criterion, is also proposed. Compared with the general minimum chordal distance criterion, the proposed quantization criterion can achieve a higher sum rate. Performance improvement of the proposed scheme is analyzed. Simulations also confirm its effectiveness.