Kashef Mohamed, Candell Richard, Lee Kang
National Institute of Standards and Technology, Gaithersburg, MD 20899, USA.
J Res Natl Inst Stand Technol. 2019 Mar 18;124:1-16. doi: 10.6028/jres.124.007. eCollection 2019.
Industrial control systems are increasingly using wireless communications to improve monitoring and control of industrial processes. In existing installations, distances and costs for installation often prohibit the running of new cables and conduits, making wireless solutions very attractive. With costs reduced, monitoring of the physical process becomes easier, and operators often desire to extend wireless to include supervisory and feedback control. Feedback control, in particular, requires certain reliability, latency, and performance guarantees that are difficult to characterize. Industrial wireless solutions rarely make quality-of-service measurements available at the control system level. When they do, indicators such as per-link packet success rate are often difficult to translate into meaningful metrics useful to the control system designer. This is especially true for multihop mesh network architectures, where it is difficult to translate link performance to system performance. In this paper, we propose a more useful method to characterize true network latency and reliability of a deployed industrial wireless network without the need for physical layer and link layer performance metrics and design knowledge.
工业控制系统越来越多地采用无线通信来改善对工业过程的监测与控制。在现有设施中,安装距离和成本常常使得铺设新电缆和管道变得不可行,这使得无线解决方案极具吸引力。随着成本的降低,对物理过程的监测变得更加容易,并且操作人员常常希望将无线扩展至包括监控和反馈控制。特别是反馈控制,需要某些难以描述特性的可靠性、延迟和性能保证。工业无线解决方案很少在控制系统层面提供服务质量测量。即便提供了,诸如每条链路的数据包成功率等指标往往难以转化为对控制系统设计者有用的有意义的度量标准。对于多跳网状网络架构而言尤其如此,在这种架构中,很难将链路性能转化为系统性能。在本文中,我们提出了一种更有用的方法,无需物理层和链路层性能指标以及设计知识,即可描述已部署的工业无线网络的真实网络延迟和可靠性。