Optimal Detector for Channels with Non-Gaussian Interference

TL;DR

This paper introduces an optimal detector for channels with non-Gaussian interference, demonstrating its superior performance over conventional methods through analytical and simulation results.

Contribution

It proposes a new optimal detection scheme for non-Gaussian interference channels, outperforming existing detectors and providing a less complex alternative.

Findings

Optimal detector outperforms conventional and SIC detectors.

The optimal detector achieves lower symbol error rates under strong interference.

Minimum-distance detector performs nearly as well as the optimal detector.

Abstract

The detection problem in the Gaussian interference channel is addressed, when transmitters employ non-Gaussian schemes designed for the single-user Gaussian channel. A structure consisting of a separate symbol-by-symbol detector and a hard decoder is considered. Given this structure, an optimal detector is presented that is compared to an interferenceunaware conventional detector, an interference-aware successive interference cancellation (SIC) detector, and a minimum-distance detector. It is demonstrated analytically and by simulation that the optimal detector outperforms both the conventional and the SIC detector, and that it attains decreasing symbol error rates even in the presence of strong interference. Moreover, the minimum-distance detector performs almost as well as the optimal detector in most scenarios and is significantly less complex.