The regressive QoE model for VoLTE

Abstract

LTE is a 3GPP full IP-based standard for wireless transmission systems. Due to its characteristic of high throughput, an efficient end-to-end QoS treatment is needed in order to guarantee good quality perceived by end users. LTE provides a native QoS-aware mechanism for end-to-end service delivering based on EPS bearer and QCI. Customers today may choose mobile services among different competing network operators according to their experienced quality of the services. The network operators must be able to react quickly to quality issues to offer more satisfied services before the users grumble. It is essential to establish a relationship between QoE and QoS reacting to the network performance. The voice service delivered by the LTE system is a kind of Voice over IP (VoIP) service with no QoS-guaranteed mechanism. LTE Standard specifies the QoS, but doesn't define the QoS/QoE relationship. This research identifies the QoS and QoE parameters, named respectively KPI and KQI for the VoLTE service, and then analyzes the QoS/QoE mathematical relationship. The main contribution of this study relies on the possibility to predict the QoE level through the functional relation with QoS. We propose the linear regression technique to derive objective voice quality metric (MOS) through subjective tests and show the correlative causality of KPIs and MOS.