On the Ice-Wine Problem: Recovering Linear Combination of Codewords over the Gaussian Multiple Access Channel
In this paper, we consider the Ice-Wine problem: Two transmitters send their messages over the Gaussian Multiple-Access Channel (MAC) and a receiver aims to recover a linear combination of codewords. The best known achievable rate-region for this problem is due to [1],[2] as $R_{i}\leq\frac{1}{2}\log\left(\frac{1}{2}+{\rm SNR}\right)$ $(i=1,2)$. In this paper, we design a novel scheme using lattice codes and show that the rate region of this problem can be improved. The main difference between our proposed scheme with known schemes in [1],[2] is that instead of recovering the sum of codewords at the decoder, a non-integer linear combination of codewords is recovered. Comparing the achievable rate-region with the outer bound, $R_{i}\leq\frac{1}{2}\log\left(1+{\rm SNR}\right)\,\,(i=1,2)$, we observe that the achievable rate for each user is partially tight. Finally, by applying our proposed scheme to the Gaussian Two Way Relay Channel (GTWRC), we show that the best rate region for this problem can be improved.