Quality Of Experience Estimation Model Of Video Streaming And Its Applications In Wireless Networks

Video streaming service in wireless networks is growing exponentially and gaining popularity, and is predicted to expose new revenue streams for mobile network operators. Although the purpose of quality of service (QoS) is to provide the optimal video streaming service, it considers quality-dependent factors limited to network or technical level which can hardly indicate the video quality perceived by the end-user. However, the success of video streaming applications in wireless networks depends on satisfying the user’s Quality of Experience (QoE), which is defined as "the overall acceptability of an application or service as perceived by the end-user". QoE is a comprehensive measurement of end-to-end service performance considering technical and non-technical parameters. Thus, it is highly desirable to be able to estimate and control the video streaming quality to meet users’QoE requirements.In order to develop an efficient no-reference estimation model for video quality estimation, we first investigate the cross-layer parameters affecting video quality, including the bit rate, frame rate and resolution at the application layer, the packet loss rate at the network layer, video content features and user equipment characteristics. Then, we present a no-reference, content-based QoE estimation model for video streaming service in wireless network. The video quality estimation model is based on radial basis function networks (RBFN) which is a kind of machine learning method with excellent approximating ability. Simulation results show that the RBFN-based QoE estimation model performs well in terms of high estimation accuracy, high Pearson correlation coefficient, low root mean square error, and small computational time.We present a QoE driven adaptation schemes for video streaming applications with RBFN-based QoE estimation model. The adaptation schemes uses packet loss rate and one-way delay trend of video packets to serve as network congestion indicators. The sender side efficiently adapts the encoding bitrate to the level of network congestion with specified adaptation strategies. Simulation results show that the bitrate adaptation scheme is an efficient control and optimization method to accurately adapt to varying network conditions and improve user’s perceived video quality.