| As information communication globalizes and generalizes,video service expands to every aspect of the modern world.It popularizes so explosively as to be a necessity in business activities of many organizations and daily life of many ordinary people.Mobile video service,in particular,is a perfect match for many people with a rapid pace of life,as soon as wireless network covers most areas with intensive human activity.The general environment of the world values and urges for all research projects regarding the effectiveness,efficiency,quality,and throughput of mobile video transmission.Nevertheless,considering that the growth in the demand for mobile video services plays a leading role in the expansion of this industry,many research questions remain unresolved and many procedures remain suboptimal in this field.The scenario in which many mobile video consumers are at close physical proximity to each other becomes an ordinary state as the population and density of video service consumers increase.When it happens,the unavoidable competitions for limited network resources cause massive and simultaneous video quality distortion,which harm the consumers' quality of experience(QoE)extensively.The novelty of this paper,however,is to transform this misfortunate scenario in the conventional sense into an advantageous setting for cooperative system design.The cooperative video transmission architecture connects the devices by initiating a local one-hop ad hoc WiFi network while keeping the existing cellular network active.The members of this connected union could defend themselves against fluctuating cellular network conditions to increase video quality stability.In addition,efficient allocations of resources and tasks in the local WiFi network could reduce within-group competitions to avoid the wireless collision.The design of this architecture faces and meets the criterion of three major challenges in this research field:First,the design should be sufficiently flexible to contribute to the transmission of most types of video.Second,its theoretical architecture should incorporate multiple devices with multiple network interface controller(NIC)on each device.Third,it should define some realistic objective functions and corresponding constraints in the optimization problems,then proposes and implements some close-to-optimal heuristic approaches when prompt analytical solutions are not achievable.To show the adaptiveness and effectiveness of the cooperative scheme,this paper presents the detailed design of the scheme when combining the idea of the cooperative architecture with Video-on-Demand(VoD),real-time and 360-degree panoramic video transmission.The contributions of each chapter of this paper are as follow:1.Mobile VoD Service on Dynamic Adaptive Streaming over HTTP(DASH).DASH is often used in scenarios such as film screening and online course viewing,which means that the quality of the video is the first order consideration of mobile video transmission,with energy reduction being the second.In light of these optimization factors,an optimization problem is proposed for a group of users,in which an objective function is defined with customized weights for the two considerations.The analytical solution of this constrained maximization problem is obtained taking advantage of the mathematical characteristics of the function.In practice,the theoretical optimum of the optimization is then approached by transforming to the key components in the architecture of a device-to-device(D2D)framework,named Green Cooperative DASH(GDASH),to promote redundancy-eliminating cooperation among devices.The cooperative scheme is tested in the real-world experiment by implementing the proposed system on an Android platform.The performance of the system on a group of smartphones shows that after controlling for the quality of the video delivered,adopting the cooperative system may save their energy by 17%on average.2.Mobile Delay-Constrained Real-time Streaming.Mobile Real-time video service is an advantageous choice for interactive mobile streaming as well as an important tool for information spread in case of emergencies and natural disasters.In either of its usage,minimizing the end-to-end delay of the video it delivers is the first priority in this type of service.However,as many clients are physically close to each other and share a cellular network to access a same real-time video,the available bandwidth allocated to each device is inevitably insufficient.In this chapter,queueing theory is exploited to model cooperative video delivery among multiple users,in order to formulate the transmission delay and goodput into a constrained optimization problem.Mathematical analysis of the problem proves the existence of an analytical optimum,and validates that the optimum is achievable through the batch gradient descent method.Based on such theoretical understanding of the problem,a cooperative scheme aiming for network congestion reduction and efficiency optimization in real-time traffic with delay constraints is proposed and named Delay-Constrained Streaming in Hybrid Networks(DC-SHN).The scheme involves a data-:flow distribution mechanism to obtain a close-to-optimal result to the optimization problem with fast convergence for efficient online operation.To test the performance of the scheme,a prototype is implemented on the Android platform using overhearing technique within the local WiFi network.The results of the experiment show that the proposed approach yields reasonably improved outcomes,including a better video peak signal-to-noise ratio(PSNR),end-to-end delay,and goodput.3.Mobile 360-degree Panoramic Video StreamingThe usage of high quality 360-degree immersive panoramic video is booming in the video industry.However,existing schemes for smartphones suffer from significant bandwidth consumption as they transmit the entire panoramic view with very high resolution.This demand for bandwidth becomes even more problematic when multiple adjacent smartphones compete to access the same content,which further challenges the capacity of the wireless network,where available bandwidth fluctuates much more than wired networks.A cooperative streaming scheme named Cooperative Panoramic Video(CoPano)is proposed for tile-based 360-degree video using Scalable Video Coding(SVC)to maximize the QoE of a group of users.An optimization problem is then formulated accordingly to choose the optimal downloading and sharing subsets from the set of all requested SVC layers of tiles,in order to maximize the effective quality of the users' viewport while meeting the feasibility of the bandwidth of heterogeneous networks.The problem is shown to be NP-hard and compose a heuristic approach.Mathematical proof is provided to show the strong similarity between the analytical solution of the constrained optimization problem and the proposed heuristic approach.In the approach,The SVC layers are ranked based on the aggregated group-level preference to guide the downloading and sharing activities of the devices.A prototype on the Android platform is developed to test the performance,and the real-world results show that our proposed scheme outperforms baseline alternatives. |
PDF Full Text Request