Performance Enhancement of QoS of Advance Wireless Networks using Machine Learning

Abstract

In advanced wireless network, it is a primarily challenged task to manage Quality of Experience (QoE) for the end-user to provide reliable and cost-effective availability of connectivity and capacity, anywhere and anytime across heterogeneous stratified network fulfilling their satisfaction from the services offered. Accurate path loss prediction of radio propagation is a prerequisite to achieve an optimum wireless system design. The growing popularity of wireless services along with a variety of emerging technologies, services, and frequency bands has kept the development of the propagation model at the center stage of research activities in the wireless industry and academia. Empirical models considered for the prediction of path loss are highly distance-dependent and do not consider other multiple parameters responsible for propagation losses. On the contrary, machine learning-based models can consider multivariate, non-parametric parameters responsible for path loss efficiently. Collected field data is used for tuning of empirical path loss models and design of machine learning path loss model. Radio propagation losses are subjected to multivariate factors like reflection, deflection, and scattering due to buildings, vegetation, hills etc. Machine learning holds to be the promising solution as it can learn from the large data sets available and builds a model through training, distinguishing the varying essential features then testing the model, and finally determines the optimal prediction. The effectiveness and efficiency of ANFIS machine learning technology motivate to apply it for predicting coverage in a complex terrain region like Uttarakhand-India. Further for mm wave wireless network a high gain array antenna system with improved antenna characteristics like reduced Side Lobe Level (SLL), stable radiation characteristics, etc. is a solution to the 5G applications. newlineThe drive test has been carried out at Dehradun, Haridwar, and Rishikesh. Through this drive test, an extensive sample of data is collected