Handoff decisions in hetnets using fuzzy logic and machine learning techniques

Abstract

Handover is a vital part of any wireless Mobile Communication Network. Efficient handover algorithms provide a gain in monetary cost effective method for enhancement in the capacity and QOS(quality of services) of cellular system. The work reported here presents a handover decision support system for wireless heterogenous communication. Fuzzy inference system along with neural network techniques in Matlab has been used for the development of the system.Results for performance of five different neural network techniques have been drawn on basis of their training time,classification accuracy and number of neurons taken. The neural tools used here are Back propagation, Radial Basis Function, Nave Bayes, Bayes net and C4.5 decision tree. On the basis of their performance, a best neural tool has been selected for the use in making handover decisions in wireless communication systems and it is seen that among all the neural network tools C 4.5 decision tree gives the better results in terms of classification accuracy and training time for handover decisions.