A novel context-aware RPL algorithm based on a triangle module operator

Abstract

For the use in low-power and lossy networks (LLNs) under complex and harsh communication conditions, the routing protocol for LLNs (RPL) standardized by the Internet Engineering Task Force is specially designed. To improve the performance of LLNs, we propose a novel context-aware RPL algorithm based on a triangle module operator (CAR-TMO). A novel composite context-aware routing metric (CA-RM) is designed, which synchronously evaluates the residual energy index, buffer occupancy ratio of a node, expected transmission count (ETX), delay, and hop count from a candidate parent to the root. CA-RM considers the residual energy index and buffer occupancy ratio of the candidate parent and its preferred parent in a recursive manner to reduce the effect of upstream parents, since farther paths are considered. CA-RM comprehensively uses the sum, mean, and standard deviation values of ETX and delay of links in a path to ensure a better performance. Moreover, in CAR-TMO, the membership function of each routing metric is designed. Then, a comprehensive membership function is constructed based on a triangle module operator, the membership function of each routing metric, and a comprehensive context-aware objective function. A novel mechanism for calculating the node rank and the mechanisms for preferred parent selection are proposed. Finally, theoretical analysis and simulation results show that CAR-TMO outperforms several state-of-the-art RPL algorithms in terms of the packet delivery ratio and energy efficiency.

摘要

低功耗有损网络路由协议 (RPL) 由因特网工程任务组设计, 主要适用于通信条件复杂、 环境恶劣的低功耗有损网络. 为进一步提高低功耗有损网络性能, 本文提出一种基于三角模算子的情景感知RPL新算法 (CAR-TMO). 首先设计了一种新的情景感知复合路由度量 (CA-RM) ; CA-RM可综合评估候选父节点的剩余能量指数、 缓存占用率、 以及该候选父节点到根节点之间路径所需的期望传输数 (ETX)、 时延和跳数. CA-RM以递归方式评估了候选父节点及其偏好父节点的剩余能量指数和缓存占用率, 以降低上游父节点对偏好父节点选择的影响. CA-RM综合使用路径上各链路ETX和时延的和值、 均值和均方差值以进一步提高网络性能. 其次, 设计了上述各路由度量的隶属度函数. 然后, 基于三角模算子和各路由度量的隶属度函数构造综合隶属度函数和情景感知目标函数 (CA-OF). 此外, 提出新的计算节点秩值和偏好父节点选择机制. 最后, 理论分析和仿真结果均表明, CAR-TMO在分组投递成功率、 能效等方面均优于RPL及其相关改进算法.