Power Allocation for Decode-and-Forward Relay in Gateway Channels

In this paper, we consider a gateway channel, in which multiple source nodes serve their destination nodes with an aid of a single relay node, and propose power allocation schemes at the relay node to maximize the sum rate. After receiving the signals from the source nodes, the relay node broadcasts the decoded messages to the destination nodes, using zero-forcing beamforming and a proper power allocation. For the power allocation, the relay node should consider not only the channel conditions to the destination nodes, but also the message rates from the source nodes, which are the destination nodes' maximum achievable rates. To develop an efficient power allocation strategy with quantized channel state information (CSI), we first assume perfect CSI at the relay node and find the optimal power allocation scheme named as
ceiled waterfilling
. Then, we modify the ceiled waterfilling taking into account the effects of the quantization errors and develop an iterative power allocation algorithm with quantized CSI. Our numerical results show that our proposed power allocation scheme considerably increases the sum rate with respect to the naive equal power allocation at the relay node in the gateway channel for all limited feedback scenarios.