Multivariate statistical approach for estimating QoE of real-time multimedia applications in vehicular ITS network

In contrast to existing systems, which inadequately evaluate real-time multimedia QoE, using Full Reference or Reduce Reference objective metric, we devised a real-time rating procedure that estimates, real-time multimedia perceived quality with no need for human participation nor reference signal.To structure the proposed QoE prediction model, we segment the multimedia/VANETs distribution network, into a framework of two quality optimization componentWe set up a procedure and heuristic to demonstrate the connection between each quality optimization component and set up a relationship between the variables of each quality Factor (i.e., the QoE parameters) as they impacted on the QoE.To deduce the collective impact of the factors on the end-user QoE, we develop a correlation model that estimates the QoE as a weighted sum of the aggregated QoE parameters Though absolute QoE assessment, requires a subjective approach, performing a subjective test to assess real-time multimedia quality is expensive in terms of time and resources. The process involved in subjective approach is not suited for assessing real-time multimedia services such as IPTV over a dynamic network such as VANETs. Thus, the only practical solution during service operation is to apply an objective quality assessment model, which produces an estimate of the perceived quality. Hence, in this paper, we propose a novel objective QoE prediction model that estimates the QoE of real-time multimedia services over VANETs. Our proposed model is based on a multivariate statistical approach, in conjunction with ordinal regression analysis, that estimates perceived multimedia service quality as a function of aggregated QoE influencing parameters. We assume that each parameter has a different weight depending on the application used. Therefore, to create a standardized QoE model, we develop a correlation model where we estimate the QoE as a weighted sum of the QoE influencing parameters. To validate the effectiveness of our proposed model, Monte Carlo simulation was carried out to investigate the model predicting capabilities. The results attest to be very promising as the proposed model exhibits good predictive ability coherent with the observed data. Display Omitted