基于车联网的多信道视频点播下载算法研究

发布时间：2018-03-11 01:35

本文选题：视频点播　切入点：车联网　出处：《华东师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着互联网的发展和移动终端的普及,视频点播(video-on-demand,简称VOD)成为了目前最受欢迎的网络应用之一。无论是在学习、工作还是娱乐中,视频点播都给人们带来了极大的便利,人们希望在旅途中也能方便地观看视频,而目前这种场景下的视频点播大多局限于使用本地缓存或者数据流量的方式,存在灵活性较差或价格昂贵等问题。车联网的出现使这些问题的解决成为了可能,用户可以通过接入车联网使用视频点播系统,然而车联网的移动性和资源有限性等特点会带来视频播放不畅、清晰度不高等问题。因此,如何设计满足各种用户需求的视频点播系统,如何利用车联网中的有限资源带来更优的用户体验成为了目前亟待解决的问题。在这种情况下,本文针对车联网中的多个场景提出了一套基于C/S架构的解决方案,能够有效满足各种用户的需求,提升用户体验。本文的主要贡献如下:·提出一套客户端视频下载框架。该框架由三个不同场景下的问题模型构成。首先,通过将3-SAT(3)问题归约到多信道视频下载问题(McPDP)证明其NP完全性。接着,通过将视频块整齐排列的多信道视频下载问题(AMcPDP)归约到最大流问题,发现其可在多项式时间内解决,并提出了两种算法来解决它。最后,对于有限存储容量的多信道视频下载问题(McDDPS),提出了两种基于贪心算法的方法来找到它的最早播放时间以及相应的下载调度方案。·提出一套服务器端信道广播协议框架。该框架包括简单分段广播协议(SSB)和区块循环广播协议(BCB)两种基于多信道广播模型的协议,对于其存储空间有限制和无限制版本,本文分别提出了下载算法,有效减小了客户端延迟。·将这两个框架扩展到多服务器的车联网环境中,提出解决多服务器环境下视频下载问题的一般步骤。该方法对多服务器环境下小型视频块和大型视频块的下载案例分别提出了解决方案,有效地提高了客户端的下载效率。最后,本文通过实验对所提出的算法进行了正确性验证及性能对比,并且对多服务器的案例进行了仿真研究。实验结果表明,本文所提出的两个框架能有效减小客户端延迟、降低存储要求,带来更好的用户体验。
[Abstract]:With the development of the Internet and the popularity of mobile terminals, video-on-demand (VOD) has become one of the most popular network applications. People want to be able to watch video easily on the road, and most of the video-on-demand in this kind of scene is limited to the way of using local cache or data traffic. Problems such as low flexibility or high price exist. The emergence of car networking makes it possible to solve these problems. Users can use video-on-demand systems through access to the car network. However, the mobility and limited resources of the vehicle network will lead to problems such as poor video playback and low clarity. Therefore, how to design a video-on-demand system to meet the needs of various users, How to make use of the limited resources in vehicle networking to bring better user experience has become an urgent problem. In this case, this paper proposes a solution based on C / S architecture for many scenarios in vehicle networking. The main contributions of this paper are as follows: 路A client-side video download framework is proposed. The framework is composed of three problem models under different scenarios. By reducing the 3-SAT3) problem to the multi-channel video download problem (McPDP), we prove its NP completeness. Then, by reducing the multi-channel video download problem with neatly arranged video blocks to the maximum stream problem, we find that it can be solved in polynomial time. And put forward two algorithms to solve it. Finally, To solve the problem of multi-channel video download with limited storage capacity, this paper proposes two methods based on greedy algorithm to find the earliest playback time and the corresponding download scheduling scheme. Finally, a set of server-side channel broadcast protocol is proposed. This framework includes two kinds of protocols based on multi-channel broadcast model: simple segmented broadcast protocol (SSBs) and block cyclic broadcast protocol (BCBB). For the limited and unlimited storage space, this paper proposes a download algorithm, which can effectively reduce the delay of the client. 路extend the two frameworks to the multi-server vehicle networking environment. The general steps to solve the problem of video download in multi-server environment are put forward, and the solutions to the downloading cases of small video blocks and large video blocks in multi-server environment are presented respectively. Finally, the correctness and performance of the proposed algorithm are verified and compared by experiments, and the multi-server cases are simulated. The experimental results show that, The two frameworks proposed in this paper can effectively reduce client latency, reduce storage requirements and bring better user experience.
【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U495;U463.67;TN948.64

【相似文献】