Evaluation of quality of experience for video streaming over dynamic spectrum access systems

In this paper we study the problem of quantifying the value of spectrum opportunities to secondary users in Dynamic Spectrum Access (DSA) systems. We especially focus on estimating the impact of key channel parameters, namely the activity patterns of the primary users on the expected quality of experience for secondary users accessing video streams over a DSA system. In our study we consider three basic types of video representing typical video content categorized according to scene change rate ranging from low-activity newscast to high-activity sports video. Through extensive simulations we show that given some information on the expected level of activity in the video, the duty cycle of the primary user alone can yield good predictors for the expected quality of experience of secondary users. Knowledge of the precise distributions of the primary user ON and OFF periods can be used to further enhance the precision of the prediction, but at least for exponential and log-normal channel access patterns the differences are rather small. Finally, we study the problem of determining the channel statistics that are needed to apply the predictor in an optimization setting. Our simulations show that high accuracy can be achieved in matter of minutes of estimation time, which is more than enough for practical deployments in typical urban environments.