智能家庭网关关键技术研究
文献类型：学位论文
| 作者 | 么刚 |
| 学位类别 | 博士 |
| 答辩日期 | 2009-05-25 |
| 授予单位 | 中国科学院声学研究所 |
| 授予地点 | 声学研究所 |
| 关键词 | 家庭网关 协议转换 IMS（IP Multimedia Subsystem） 机器学习 QoS（Quality of Service） |
| 其他题名 | Studies on Key Technologies of Intelligent Home Gateway |
| 学位专业 | 信号与信息处理 |
| 中文摘要 | 随着家庭网络技术的发展，人们对于居住环境的健康性、安全性、舒适性和便捷性的要求越来越高。而IMS网络技术、人工智能技术、多媒体技术等和家庭网络技术相互融合，家庭网络的发展表现出了新的方向：1）互操作性：使家庭网络和IMS设备互通互操作，用户可以随时随地的享受家庭网络和移动网络中服务。2）学习性：使家庭设备自动学习家庭网络用户习惯，使设备工作符合用户习惯。3）高QoS保障：使家庭网络QoS满足基于QoS的业务分类、流媒体的流量突发性等问题。 本文依托国家863项目《家庭网络与IP多媒体子系统网络融合中的关键技术研究》，对智能家庭网关的关键技术展开研究，以期解决家庭网络学习性问题、QoS问题、家庭网络与IMS网络设备互操作等问题。本文的主要贡献和创新点如下： 1）给出一种智能家庭网关的系统设计。作为家庭网络的接入中心和控制中心，智能家庭网关具备有线/无线接入点、路由、策略控制、SIP信令代理，协议转换等功能，以此为核心设备可以实现家庭网络的设备互操作性、学习性以及QoS保障。 2）提出一种IMS/Home network协议转换方法。该方法将家庭网络协议所共有的抽象设备行为和IMS（IP Meida Subsystem）网络中SIP（Session Initiation Protocol）协议信令集进行映射，实现家庭网络和IMS网络的协议转换，进而实现家庭网络设备和IMS网络设备的互操作性。该协议转换方法兼容现有的主流家庭网络协议，并对现有的家庭网络设备和IMS设备具有透明性。 3）提出一种以智能家庭网关为核心设备、全局和设备两级代理的家庭网络学习模型。全局代理维护整个家庭网络的状态，学习家庭网络用户的生活习惯并决定家庭网络的服务该由哪个设备提供以及何时提供；设备代理学习与本设备相关的联动知识，并提供具体服务。该模式可使家庭网络用户获得个性化的服务。 4）提出了一种家庭网络和IMS网络融合下的基于智能家庭网关的QoS策 略模型。在信令层面，该模型综合考虑家庭网络策略库和IMS策略库做出决策；在数据层面，该模型的策略执行点改进了资源预留、准入控制、流量整形、缓存管理、队列调度等算法，较好的解决了IMS网络中的业务分类以及媒体流式服务的流量突发等问题。 |
| 英文摘要 | With the development of the intelligent home network, people request more and more for the health, safety, comfort and conveniency of the living environment. The IMS network technology, the artificial intelligence technology and the multimedia technology merge with the intelligent home network technology. There are some new directions of the home network: 1) interoperability: the devices in the home network and in the IMS network can touch with and control each other, furthermore the users can enjoy the services anytime, anywhere. 2) learning functions of the devices: the devices should learn the living habit of the users by themselves and adjust their work. 3) good QoS protection: the SLS(Service Level Specification) technology in the IMS network field and the high instantaneous bit rate problem in the VBR (variable bit rate) streams in the multimedia field all request good QoS protection in the home network while these technologys merge with the home network. Based on the Hi-tech Research and Development Progran of China（863）”Study on Key Technologies of the MFC for home network and IMS”, this thesis researches the tendency of the intelligent home network and studies the mobility problem, the learning problem and the QoS promlem. The main contributions are as follows: 1) This thesis gives the design of the intelligent home gateway. The gateway is the access center and control center of the home network, which has rich functions such as cable access, wireless access, routing, policy control, SIP signaling agent, protocol conversion and so on. The home gateway can realize the interoperability of the deivces, learning functions of the devices and the QoS protection. 2) A method for the protocol conversion of the home network and the IMS network is given. The method maps the the abstract behavior of the home devices to the SIP protocol messages, realizes the protocol conversion and the interoperability of the devices. Futhermore, the method is compatible with the main home network protocols and is unified and transparent for the home network devices and the IMS deivces. 3) A deivce learning model based multi-level agents is presented. In this model the home gateway is the core equipment. The model is composed of overall agent and device agents. The overall agent summarizes the rules of service through the long-term data of the family and guides the work of equipment agents; the equipment agents learn automatically, detect the change of the family environment and make optimal choices. The model makes users in the home network enjoy the personalized services. 4) A policy based QoS model in the home gateway is presented while the home network and the IMS network converge. The model is composed of the PDF and the PEP. In the signaling level, the model considers the policies in the IMS network and the policies in the home network. And in the data level the PEP improves algorithms such as admition control, RSVP, buffer management, queuing and scheduling and traffic shaping to satisfy the new QoS problem such as business category in the IMS network and the sudden flow of the multimedia stream. |
| 语种 | 中文 |
| 公开日期 | 2011-05-07 |
| 页码 | 135 |
| 源URL | [http://159.226.59.140/handle/311008/520]  |
| 专题 | 声学研究所_声学所博硕士学位论文_1981-2009博硕士学位论文 |
| 推荐引用方式 GB/T 7714 | 么刚. 智能家庭网关关键技术研究[D]. 声学研究所. 中国科学院声学研究所. 2009. |
入库方式： OAI收割
来源：声学研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。