An Autonomous Deterioration Prediction Based Handover Model for 5G Networks

Abstract

Millimeter Waves (mmWaves) are very sensitive to the user and topographic dynamics. They adversely leads to irregular cell patterns that ultimately affect connectivity and access quality in mobile networks. For instance, following the adoption of mmWaves in fifth-generation (5G) networks, irregularities in mmWave cell patterns subject most classic Handover (HO) schemes to wrong, too early, or late HOs scenes. To remedy the HO challenges and choose more reliable links, this paper proposes a HO scheme that predicts and assesses the target link's immediate behaviors and deterioration pattern post-HOs. This work uses the Jump Markov Linear System (JMLS) model, which takes into account abrupt changes in the system dynamics. Thus, this work exploits the JMLS capability to account for likely abrupt changes as users switch between Line of Sight (LOS) and non-NLOS scenes following user/topographic dynamics. Simulation results show that knowledge about link deterioration patterns rather than immediate behaviour after HO helps cut unnecessary HO failures.