基于业务感知的矿山物联网自治QoS保障机制研究

Mine internet of things provides a new idea for intelligent safe production, warning and rescue in mine, but QoS of the large flow and multi-service mine internet of things is an urgent problem to be solved. Aiming at service adaptive demand of mine internet of things and security problem of network terminal QoS in coal mine under bad environment. The project uses cross-layer cognitive design by cognitive element, and introduces autonomous control loop technology to establish autonomous QoS architecture model of mine internet of things which can meet the service needs of mine internet of things. Ontology is introduced to describe the parameters characteristic of service by semantic，the rough set theory and semantic distance method are used to research the intelligent service-aware of multi-dimensional and multi-granularity. this project uses vague set to describe the parameters of QoS, and puts forward much QoS attribute fusion method based on the theory of evidence ER and dynamic self configuration of QoS for adaptive adjustment of QoS priority. In order to explore the mechanism of adaptive services providing of mine internet of things based on service-aware,this project uses end road coordination control strategy which based on level of service source control and distributed cognition link control, adopts active queue management methods based on fractional PID for QoS adaptive control of mine internet of things. It is foundation for the realization of sensor mine to establish autonomous QoS system of mine internet of things based on service-aware for safety production of mine.

矿山物联网技术为煤矿安全生产与预警救援新体系提供了新思路，但大流量多业务矿山物联网的服务质量保证是亟待解决的问题。针对矿山物联网服务提供的自适应需求及煤矿恶劣环境下的网络终端QoS保障问题，项目采用跨层认知设计思想，融入认知元素，引入自治控制环技术，建立满足矿山物联网业务需求的QoS体系结构模型；引入本体论对业务特征参数进行语义描述，利用粗糙集理论和语义距离方法进行多维、多粒度网络业务感知；采用Vague集描述QoS参数，提出基于证据理论ER算法的Vague集多QoS属性融合方法与QoS动态自配置机制，进行终端业务QoS优先级动态自适应调整；采用业务源端控制层面和分布式认知链路控制层面的端路协同控制策略，结合分数阶PID的主动队列管理方法进行矿山物联网QoS自适应控制，探索基于业务感知的矿山物联网服务QoS保障机制。建立基于业务感知的矿山物联网自治QoS系统，为实现感知矿山奠定基础。

矿山物联网技术为煤矿安全生产与预警救援新体系提供了新思路，但物联网中的一些关键业务，如测距、多元数据融合、位置感知等业务的矿山物联网服务质量保证是亟待解决的问题。针对这些业务的时间同步、延时控制等关键QoS保障问题，项目深入研究满足多业务条件下QoS保障的分层的自治网络拓扑模型，基于层次化网络拓扑控制方法，结合移动Agent，采用分簇机制进行拓扑控制，利用移动Agent可以独立的感知环境的变化,并快速的做出响应，使网络保持在最优状态。深入研究网络QoS延时预测模型和延时评估方法，进而基于网络延时评估模型进行网络QoS延时控制，采用基于ＲＢＦ神经网络的智能PID控制器进行网络延时控制，并通过ＲＢＦ神经网络动态调整ＰＩＤ控制器参数，实现ＰＩＤ控制器参数自整定，提高网络延时的控制精度和稳定性。基于智能蚁群算法研究多业务条件下QoS组播路由问题，在分布的网络中寻找一条具有足够资源的最优路径来满足端到端服务质量。基于分层的自治网络拓扑，项目从同步消息量、同步周期和占空比角度研究低能耗同步算法，实现群内节点持续同步，群间节点按需同步，为不同业务的同步QoS需求提供技术支撑。项目基于矿山物联网服务关键参数研究QoS保障机制，建立基于业务感知的矿山物联网自治QoS系统，为实现感知矿山奠定基础。