Two-Dimensional Information Preamble Multiplicity Detection in Cell-Free Massive MIMO Systems

In the grant-free random access (GFRA) scenario, orthogonal preamble collisions of massive active user equipments (UEs) may lead to communication access failure. However, cell-free massive multiple-input multiple-output (mMIMO) system has good signal spatial sparsity, which can largely mitigate the interference from colliding UEs in the preamble domain. Therefore, blind detection of the preamble collision situation before formal access can provide relevant information about the preamble collision for subsequent access. This paper proposes a fully-connected neural network (FNN) based preamble multiplicity detection, specifically, to obtain information about the number of colliding UEs choosing the same preamble and the APs near the collision UEs based on the preamble energy and angle of arrival (AOA) information of the received signal. The simulation proves that preamble multiplicity detection using multidimensional information can yield reliable predictions that are significantly better than the results of detection using only one-dimensional information.