Stochastic Model for Intelligent Cell Level Mobility Prediction in Next Generation Wireless Networks

Abstract

Next generation wireless networks (NGWN) are being designed to support very high data rate, very less delay & jitter, seamless movement accross operators and geographical regions with permissible faster speed and many more quality of services (QoS). More frequency spectrum in the medium to high band range has been allotted to meet the desired QoS in the 5G wireless networks. This high frequency signals have smaller length and penetration causing dense deployment of smaller cells for wide coverage in NGWN. Smaller cells means more frequent handover of users between cells. This change of connecting cells, i.e. mobility of mobile users, is a matter of great concern to the service providers for resource management and maintaing desired QoS. In this article, a stochastic model based on connected mobile population in a base station for mobility prediction has been proposed to impart machine intelligence. The random motion of mobile devices and their connection status with an access point (AP) or base station (BS), also known as cell, at different time interval of operation is studied. The transition probabilities of a BS required to accommodate the handoff request of a mobile device at an interval is estimated from the BS records and a Markov model based mobility prediction method proposed. The proposed prediction method does not add any traffic overhead for collecting data. It predicts the number of handoff and fresh connection requests to serve at an interval and can fecilitate resource reservation, conjestion control and smooth handoff. Some practical application senarios of mobility prediction are also discussted. The article also highlights the present open challenges and potential future research issues in this domain.