Enhancement of QoS using Optimization of Communication Parameters in Vehicular Adhoc Networks

Abstract

Vehicular Ad hoc Networks (VANETs) are attaining vital focus from research community as well as from manufacturers. All major vehicle manufacturers have proclaimed to embrace VANETs ability in their products. Fine tuning the parameters manually by the network operator or administrator is quite time consuming. So this claims the need for having a good optimal technique that could solve the issue and improve the quality of service in a better way. Primarily, the main purpose of these networks is to provide safety related services. Then, now-a-days investigators are working to provide streaming services in VANETs e.g. video call, online gaming, video conference etc. Importance of these applications requires not only just bandwidth but focuses on less delay, such strict Quality of Service (QoS) requirements are to be addressed in order to adapt the users need. Researchers are addressing this dispute, as it will become a vast problem when most of the vehicles will have VANETs capabilities. A novel technique for optimization of QoS in Vehicular Ad hoc Networks, which will cater for the stringent requirements of the future vehicular networks is proposed in this research. The proposed Refined Regression Statistical Classifier Model (RRSCM) aims to resolve the above speculated issue and aids in improving the QoS. The optimized results compared with the existing built in classifier models has revealed the overpowering performance of proposed RRSCM. Artificial Neural Networks classification seems to be quite time consuming and the accuracy percentage rate of RRSCM outperforms ANN. The proposed multivariate analysis scheme optimizes by parting parameters and predicts the traffic behavior throughput. It aids to improve the QoS and envisage the traffic over VANETs. Earlier, the parameters were as such used as predictor variables in multiple linear regression (MLR) models. The other classification models used for prediction are evaluated with independent traffic samples. The root-mean-square percent relative error (RMS%RE) will measure the merit and characterize the performance of various other classification models over the proposed RRSCM. This RRSCM technique, when adopted by the cellular network operators and ISPs, shall be benefited and will be able to provide better services to the vehicles on the road.