The Research And Application Of Adaption Method For DASH-based Mobile Streaming

With the rapid development of mobile Internet,the continuous hardware performance improvement and explosive popularity of mobile device,the demand for video on mobile devices has been increasing.Business of mobile streaming media has been developing rapidly.MPEG-DASH standard can provide good compatibility and flexibility for adaptive media streaming technology.Therefore,the research of mobile streaming media system has a good prospect in market applications.It is of great practical significance to save network resources and satisfy the high-quality video experience requirements of users..This paper proposes a client-based rate adaptive mechanism and a user QoE evaluation method.The research result is applied in a mobile streaming system based on MPEG-DADH protocol.The main work and achievements are as follows:(1)Buffer-based rate adaptation method and oscillation compensation mechanism for multiple competing DASH-based clientsIn this paper,we present buffer based adaptation logic on the client which uses the cache partition strategy combined with logarithm model to calculate the video bitrate.Then an adaptive smoothing function is applied to increase the rate switching stability.An improved Logistic model is used to affect smoothing factor dynamically according to the buffer filling status.The effective oscillation compensation mechanism is added for adaptation decisions due to rate fluctuations.Finally,from the perspectives of stability and bandwidth efficiency,multiple simulation experiments are presented to evaluate the performance of our algorithm.(2)Sensor-based adaptive rate optimization in mobile media streaming environmentThe streaming media on mobile devices occupy a large amount of network traffic.Wireless networks are mostly shared which may result very tight in the distribution of network bandwidth resources.This paper designs a sensor-based hybrid rate decision model based on user's viewing position,user's interest and equipment state to optimize the buffer-based rate adaption mechanism of multiple clients.The experimental results show that the scheme can effectively save the bandwidth resource of wireless network.(3)The QoE evaluation model for mobile streaming mediaThe existing user QoE model of mobile streaming lacks considering users' psychology and the environment of mobile device.This paper designs a user QoE evaluation method based on psychology.The weight distribution of the whole media streaming is assigned according to serial position effect of psychology.Then the user experience quality is evaluated synthetically by using the rate level of the streaming segment,rate switching status,device state and user'viewing position.Finally,the linear regression method is applied to analyze the results of the QoE rating.The QoE evaluation model reflects the real QoE experience of users.(4)MPEG-DASH based streaming media applicationBased on the above research results,a DASH-based mobile streaming on-demand system is designed.It combines the multi-clients rate adaptive mechanism and sensor-based adaptive rate optimization method.The QoE model based on psychology can feedback user experience.