Pattern Based Mobility Management in 5G Networks With a Game Theoretic-Jump Markov Linear System Approach

The fifth generation (5G) mobile communication adopted the usage of Millimeter Wave (mmWave) bands to ignite prospects of gigabit data rates in mobile networks. However, mmWave propagation is highly susceptible to competing factors of user and topographic dynamics: they formulate irregular cell patterns. The irregularities in mmWave cell patterns cause unreliable connectivity and can instigate unnecessary Handoffs (HOs). This behavior ultimately increases the risk of 5G link failures. To improve mmWave link connectivity hence guarantee continuous connectivity in 5G mobile communication, this paper proposes a HO scheme that predicts target link deterioration patterns to select the most reliable mmWave link for a mobile user. The scheme is based on Game Theory (GT) and Jump Markov Linear Systems (JMLS). JMLSs are known to account for abrupt/erratic changes in system dynamic predictions. We amalgamate GT with JMLS capability to predict target mmWave link pattern/behavior after the HO execution. Specifically, given channel gain and received power variation over distance, the GT-JMLS HO scheme predicts the sustainability of the signal-interference-noise ratio (SINR) pattern of a target link above threshold. This is paramount to reducing the selection of mmWave links that prematurely fail or require multiple HOs to sustain connectivity over a short distance or period. Our simulation results show that our proposed HO scheme offers target links with higher: throughput, energy efficiency, reliability, and longer dwell time between HOs than classical HO schemes.