A Neural Network Based Handover Management Strategy for Heterogeneous Networks

Abstract

One of the key challenges for improvement of quality of services (QoS) in Heterogeneous wireless networks is the design of Vertical Handover (VHO) Management strategy. VHO is required to guide the decision for a mobile terminal (MT) to handoff between different types of networks. This is an essential task to cope with various multimedia services QoS settings. In this paper, we present a machine learning scheme based on Neural Network for calls vertical handover in heterogeneous networks. The Neural Network Based Handover Management Scheme (NNBHMS) of this paper aims toward achieving seamless connectivity and Always Best Connected (ABC) call status for group mobility over a set of heterogeneous networks. The proposed scheme evaluates and creates relationships between different decision criteria related to heterogeneous networks conditions, terminal capabilities, application requirements, and user preferences. Afterward, the estimates of each attribute are forwarded to neural network to select the optimal access network. The proposed scheme is applied for vertical handover Management in heterogeneous networks offering both real time services (voice over IP services), and data Services (packet data traffic). Through the implementation of neural networks based machine learning approach, the proposed research scheme allows solving the complexity of the handover decision process resulting from the multitude dimensions of the decision criteria and the dynamicity of many of its components. The performance results evaluated through simulation show that the use of the a neural network based machine learning scheme to carry out the Handover process can enhance the QoS perceived by both types of voice and data service while fulfilling to great extent the user preference.