End-to-End Multimedia Quality Estimation with Robust Optimization in Real-World Mobile Computing Networks

Abstract

Quality of experience (QoE) parameters describe the end-to-end (E2E) quality as experienced by the mobile users. These are difficult to measure and quantify. On the one hand, system quality of service (SQoS) parameters are metrics that are close related to the network status, and defined from the viewpoint of the service provider rather than the service user. On the other hand, E2E service quality of service (ESQoS) parameters describe the QoS of the services and they are obtained directly from the QoE parameters by mapping them into parameters more relevant to network operators, service providers and mobile users. A useful technique for mobile network planning and optimization is to build quality estimation models for mobile voice and video telephony service. Our research is focused on developing statistical estimation models extracted by measurements acquired via a drive-test measurement campaign of a commercial UMTS multimedia network. Regression estimates, computed with a robust optimization strategy, suggest a weaker dependence between the SQoS and ESQoS parameters and connect the strength of the dependence with the accuracy of the measurements used to compute the estimates.