面向工业需求的低功耗有损网络路由协议研究

发布时间：2018-03-19 12:15

本文选题：低功耗有损网络　切入点：RPL　出处：《浙江理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：针对低功耗有损网络的特殊路由需求及四个主要方向的应用场景,互联网工程任务组下的ROLL组拟定了RPL路由协议,RPL协议适用于其中的工业自动化应用场景。目前WSN适用于工业自动化方向,市场占有率最高的为ISA100.11a,其采用6Lo WPAN连通网络,6LoWPAN采用RPL作为路由协议。RPL协议是基于IPV6架构,通过构建DODAG形成网络拓扑,DODAG的构建依赖目标函数OF所采用的路由度量和路由约束。OF0和MRHOF是RPL中两类主要的目标函数,其中OF0采用距离根节点相对位置最近的点作为选择最优父节点的标准,路由度量单一,这加剧了各节点的负载及能耗不均衡,导致部分节点丢包或能量提前耗尽。在LLN环境中,节点不均匀的能耗将会显著减少网络的覆盖和寿命。路由回路也是RPL协议亟待解决的问题。路由回路会影响网络收敛,导致数据包投递率下降,并大量占用链路带宽。RPL协议中,当链路质量等级LQL发生变化或网络拓扑结构改变时,极易发生路由回路。RPL中典型的回路有DODAG回路和DAO回路。RPL采用的回路避免规则为:max_depth规则和拒绝贪婪规则,但仍无法避免路由回路的产生。本文从RPL的工业应用场景需求出发,以能耗和可靠性为研究重点,相应的针对RPL协议的目标函数OF0和回路避免两方面进行优化,主要做了以下工作:(1)为解决目标函数OF0路由度量单一导致的节点负载及能耗不均衡的问题,本文提出采用综合的路由度量Metric(R,E),即从MC度量容器内选取能量度量,通过设置节点能量格式与计算节点能耗引入能量,研究将能量度量和OF0原本的路由度量相融合。使OF0中的节点在选取父节点时,不再是单一的对比Rank值,而是综合考虑节点Rank与剩余能量。(2)为提升优化性能,本文提出两种使用Metric(R,E)优化OF0的方案:方案一在原始协议基础上对比节点剩余能量;方案二则在Rank1"gRank2时对比剩余能量,剩余能量相等及Rank1Rank2时,由Rank值对比决定最优父节点。在Contiki系统中,修改RPL协议中OF0相关代码。通过COOJA仿真,对比优化前后能耗和相应节点数据包转发量,验证了优化方案初步解决了网络中节点能耗差异较大的问题,且方案二优化性能最佳。在优化前论证了SINK节点置于圆心位置时,能耗最低,在此基础上优化,更加保证节点低能耗。在采用方案二时,对比了单SINK节点、双SINK节点与是否采用优化OFO方案的网络整体能耗差异,证明在单SINK节点下OF0优化方案,优化前后网络总能耗差别最为显著。(3)为解决路由回路问题,做到深度回路避免,在RPL拒绝贪婪及max_depth规则的基础上,本文提出采用基于Rank门限的候选父节点筛选机制。在DODAG中,Rank相等的点可能引起路由回路;在链路变化时,DIO消息及节点信息更新滞后,导致节点和子节点之间互相选取作为父节点,从而出现回路。将这两类容易导致回路的节点从候选父节点集合中筛除。依此修改协议代码后,通过对比100节点和160节点环境下的数据包投递成功率和平均端到端时延,验证了优化方案部分解决了路由回路问题。
[Abstract]:Application scenarios for low power loss with special routing requirements of network and the four main directions, the Internet Engineering Task Force under the ROLL group to the RPL routing protocol, RPL protocol applies to the scene of industrial automation applications. Currently WSN is suitable for the direction of industrial automation, the highest market share for ISA100.11a, which uses 6Lo WPAN 6LoWPAN connected network, using RPL as the routing protocol.RPL is based on IPV6 architecture, the network topology is formed by constructing DODAG, DODAG is built upon objective function used by OF routing metrics and routing constraints.OF0 and MRHOF are the two main types of objective function RPL, which OF0 uses the relative position of distance from the root node as the nearest point selection the optimal parent node routing metric standard, single, which exacerbated the load and energy consumption of each node is not balanced, resulting in part of packet loss or energy depletion in the LLN environment. In energy consumption will significantly reduce the non-uniform node coverage and network lifetime. RPL protocol is a routing loop problem to be solved. The routing loop will affect the network convergence, resulting in packet delivery rate, and take up a lot of bandwidth in the.RPL protocol, when the change of network topology or link quality grade LQL, very easy loop typical routing loop in.RPL DODAG and DAO.RPL using the loop loop loop: max_depth rules and rules to avoid rejection greedy rules, but still can not avoid the routing loop. This paper from the industry application requirement of RPL based on energy consumption and reliability as the research focus, the corresponding RPL protocol the objective function of OF0 and circuit to avoid the two aspects of optimization, mainly do the following work: (1) to solve the objective function OF0 node routing metric caused by a single load and energy consumption is not balanced The problem is proposed in this paper, using an integrated routing metric Metric (R, E), from the MC metric energy container selection metric, by setting the node energy format and calculation of node energy consumption into energy, study the energy metric and the original OF0 routing metric integration. The OF0 in the selected node in the parent node. Comparison of Rank is no longer single value, but considering the node residual energy and Rank. (2) to improve the optimization performance, this paper puts forward two kinds of Metric (R, E) OF0 optimization solutions, a comparison in the original agreement based on the residual energy of nodes; scheme two in Rank1 gRank2 compared the residual energy the residual energy, equal and Rank1Rank2, by the comparison of Rank value determines the optimal parent node. In the Contiki system, modify the RPL protocol in OF0 code. Through COOJA simulation, compared before and after optimization of energy consumption and the corresponding node packet forwarding, verified optimization The case initially solved the problem of the energy consumption of network node difference scheme, and the best performance. Two optimization demonstrates the optimization SINK node at the center position, the lowest energy consumption, the optimization on the basis of this, to ensure low energy consumption. In the scheme of node two, compared to the single SINK node, SINK node with double the overall energy consumption of the network optimization scheme of OFO difference that OF0 optimization scheme in single SINK node, before and after optimization of the total energy consumption of the network. The most significant difference (3) in order to solve the routing loop problem, do the depth of circuit based on RPL to avoid greed and max_depth rules, this paper adopts the candidate parent node selection mechanism Rank based on the threshold. In DODAG, Rank equal point may cause a routing loop; in the link changes, DIO messages and node information update lag, resulting in between nodes and sub nodes each other is selected as the parent node, Thus the circuit. The two kinds of easily lead to circuit nodes from the set of candidate parent node deletion. According to the protocol code after the delivery success rate and average end-to-end delay compared with 100 nodes and 160 nodes under the data packet, to verify the optimization scheme to solve the routing loop problem. Part

【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN915.04

【参考文献】