Prediction model for offloading in vehicular Wi-Fi network

Abstract

It cannot be denied that, the inescapable diffusion of smartphones, tablets and other vehicular network applications with diverse networking and multimedia capabilities, and the associated blooming of all kinds of data-hungry multimedia services that passengers normally used while traveling exert a big challenge to cellular infrastructure operators. Wireless fidelity (Wi-Fi) as well as fourth generation long term evolution advanced (4G LTE-A) network are widely available today, Wi-Fi could be used by the vehicle users to relieve 4G LTE-A networks. Though, using IEE802.11 Wi-Fi AP to offload 4G LTE-A network for moving vehicle is a challenging task since it only covers short distance and not well deployed to cover all the roads. Several studies have proposed the offloading techniques based on predicted available APs for making offload decision. However, most of the proposed prediction mechanisms are only based on historical connection pattern. This work proposed a prediction model which utilized historical connection pattern, vehicular movement, and driver profile to predict the next available AP. The proposed model is compared with the existing models to evaluate its practicability.