NETWORK TRAFFIC PREDICTION IN 4G NETWORK BASED ON NEURAL NETWORK FOR IMPROVING QOS
K. Sashidhar Reddy, S. Varadarajan · 2014
Many mechanisms based on bandwidth reservation and handoff mechanism is proposed in the literature to decrease the connection dropping probability for ha ndoffs in mobile 4G communications. The handoff events occur at higher rate in packet-switched four th generation mobile communication networks than in traditional cellular systems. Hence, an efficient b andwidth reservation mechanism for the neighboring cells is critical in the process of handoff during the co nnection of multimedia calls. This mechanism avoids the undesired force termination and waste of limited ba ndwidth in fourth generation mobile communication networks, particularly when the mobility of the mob ility device is high. In this paper, a phased solut ion of priority detection, mobility scheduling and effecti ve bandwidth estimation have been proposed. The motivation of the phases is to provide the Quality of Service (QoS) at regular communication and high mobility condition by considering the physical para meters. Meanwhile, a neural network based servic e model is incorporated to accommodate new metrics used in processing handoffs and task scheduling in fourth generation networks. This mechanism provides the advantages in terms of choosing the tasks in a priority based scenario and providing un-interrupte d service at the time of handoffs as well as an eff ective way of utilizing bandwidth.