Active user detection is a standard problem that concernsmany applications using random access channels in cellular orad hocnetworks. Despite being known for a long time, such a detection prob-lem is complex, and standard algorithms for blind detection have to tradebetween high computational complexity and detection error probability.Traditional algorithms rely on various theoretical frameworks, includingcompressive sensing and bayesian detection, and lead to iterative algo-rithms, e.g. orthogonal matching pursuit (OMP). However, none of thesealgorithms have been proven to achieve optimal performance.This paper proposes a deep learning based algorithm (NN-MAP) able toimprove on the performance of state-of-the-art algorithm while reducingdetection time, with a codebook known at training time.
Authors: Cyrille Morin (Inria), Diane Duchemin (Inria), Claire GOURSAUD (CITI INRIA), Jean-Marie Gorce (CITI INSA-Lyon), Leonardo Cardoso (Insa-Lyon),
Hide Authors & Abstract