Deep Learning-Based Algorithm for Optimizing Relay User Equipment Activation in 5G Cellular Networks

This article addresses the problem of optimally using the relay capabilities of user equipment (UE) to augment the radio access network (RAN) in 5G deployments and beyond. This can be particularly useful in coverage constrained scenarios, such as those using millimeter waves, due to the difficulty radio signals penetrate some structures. This can lead to signal blockages and high penetration losses when providing outdoor-to-indoor coverage. To overcome these limitations, the use of relay UEs (RUEs) is seen as a possible solution to effectively extend the coverage of a cellular network. In this context, this article proposes a deep learning-based algorithm to optimize the decision regarding when RUEs should be activated and deactivated in accordance with the benefits they can provide for increasing the spectral efficiency and decreasing outage probability for the network users. The obtained results reveal a promising capability of the proposed solution to activate the most beneficial RUEs given the network conditions being experienced, leading to improvements of average spectral efficiency of 12.3% and reductions of outage probability of 89% with respect to the case without relays.