Soft handoff evaluation and efficient access network selection in next generation cellular systems

Abstract

The increased motivation (by service providers) to offer user-centric and seamless communication services - that satisfies users' quality of experience (QoE), has manifested a myriad of challenges in the field of wireless communication; and given the increased traffic capacity and sudden explosion of cellular devices, communication systems are constantly threatened by performance related issues - including soft handoff. Although intelligent techniques have evolved to provide solutions to these issues, they are yet to flourish in the area of soft handoff. This contribution therefore proposes a framework that integrates two components: (i) machine learning methodologies: self-organizing map (SOM) and pattern classification - for robust performance evaluation of available soft handoff data; (ii) multiple attribute decision making mechanisms (MADM): the Analytical Hierarchy Process (AHP) - which result feeds the Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS) - for efficient access network selection. Implementation of component one of the design revealed that SOM enabled a precise visualization of handoff features that influenced the system performance; and the error levels of training, validation and test dataset, with number and percentage of correct and incorrect classifications, were obtained from our pattern classifier. Implementation of component two of the design for four heterogeneous (access) networks indicated that although network two (N2) was selected as best access network by TOPSIS and network three (N3) by Synthetic Extent Analysis (SEA) - a method adopted in a related paper, for a particular application; both TOPSIS and SEA selected N1 as second best alternative access network and network four (N4) as third best alternative network, despite the issue of ranking abnormality in TOPSIS. Further, AHP and TOPSIS can effectively be applied as MADM algorithms in handoff decision framework for selecting the best available network for handoff.