Performance Improvement of D2D Random Access in 5G Communications

The development of fifth generation mobile networks has laid the ground for the creation of new applicative scenarios, characterized by increasingly demanding requirements. In particular, to satisfy low latency requirements, an aspect to take into account is represented by the random access procedure, performed by devices for initial connection to the network or to switch to a connected state afterwards. Because of the limited amount of resources available for random access, the presence of a high number of devices requesting access at the same time can generate congestion and require multiple access attempts, increasing latency and compromising the fulfillment of the requirements. The problem was already addressed in the scientific literature, with the proposal of a series of solutions to improve the random access procedure. This thesis work aims at enriching these efforts, by leveraging device-to-device communications and the reservation of random access resources, and by proposing a dynamic strategy to alleviate performance deterioration at runtime and intervene on the system with reconfiguration operations, in order to solve possible congestions. The studied scenarios are analyzed by means of the ns-3 network simulator, extended by the additional mmWave and MilliCar modules, for the simulation of 5G networks and device-to-device communications through the cellular sidelink. The results obtained from the simulations have the objective of demonstrating the advantages introduced by the dynamic reconfiguration strategy, and of gaining useful insights on the usage of the proposed enhancements for the random access procedure.