On the Achievability of Interference Channel Coding

TL;DR

This paper demonstrates that interference channel coding can be achieved without rate-splitting, using sparse matrix codes and constrained random generators, and extends known inner regions.

Contribution

It shows rate-splitting is unnecessary for certain inner regions and introduces new code constructions with sparse matrices.

Findings

Rate-splitting is unnecessary for achieving Han-Kobayashi and Jian-Xin-Garg regions.

Codes using sparse matrices with logarithmic column degree are effective.

Possible extensions of known inner regions are established.

Abstract

This paper investigates the achievability of the interference channel coding. It is clarified that the rate-splitting technique is unnecessary to achieve Han-Kobayashi and Jian-Xin-Garg inner regions. Codes are constructed by using sparse matrices (with logarithmic column degree) and the constrained-random-number generators. By extending the problem, we can establish a possible extension of known inner regions.