Abstract

Providing an effective quality of service (QoS) can be very challenging, especially in mobile or dynamic environments. The goal of this study is to improve the Internet in a way that anticipates user mobility allowing for more efficient and responsive resource allocation through real-time resource management and connectivity prognostic. The ability for the model to determine typical paths, times, and transition probabilities between access points is accomplished by exploring historical location data, mobility traces, and user activity in a networked environment. The proposed findings can be integrated into network control algorithms requiring spatial predictive data for future user behaviours such as prefetching, intelligent handover, and load balance to configurable infrastructures. User mobility predictions of future mobility are augmented by the combination of machine learning and Markov Chains. Usability testing conducted utilizing real-world mobility datasets is suggestive of many improvements in terms of connection stability, reduction of latency, and increased efficiency of bandwidth utilization. The supporting evidence obtained from this study is supportive of the hypothesis that network management which aware of the mobile user will enhance performance and experience in urban areas and smart cities. This research is important because it further develops a notion of intelligent, human-centric communication systems towards 5G (and beyond) to reframe spatial and temporal user behaviour to be responsive, anticipatory internet service structures and/or systems.