Adaptive feedback compression for joint transmission coordinated multi-point

In this paper, we study adaptive feedback compression of explicit multi-cell channel state information (CSI) for joint transmission coordinated multi-point (JT CoMP) in future mobile networks. The scheme comprises of three steps: First, the strongest cells are clustered using a network-defined threshold. Next, the strongest taps are selected using a threshold related to noise and out-of-cluster interference. Finally, adaptive quantization is applied so that the channel mean square error (MSE) remains sufficiently small. We show that the number of quantization bits grows linearly with the signal to noise ratio (SNR) in dB, as required by information theory. Results indicate that explicit CSI from all relevant cells around a terminal can be provided using few kbits per reporting interval in 20 MHz bandwidth. The scheme is implemented and tested in real time.