面向农田物联网复杂事件处理的时空事件模型

Time and Space Event Model for Complex Event Processing in Internet of Things in Farmland

在农田物联网中,控制命令往往由具有多维度信息的复杂事件触发,因此需要根据传感器所监测的大量单一事件检测复杂事件,即复杂事件处理。复杂事件模型描述了原子事件组合成复杂事件的组合模式,是复杂事件处理的基础。现有的复杂事件模型主要考虑原子事件的时间分布特性,而未考虑农田事件具有的空间分布特性。本文研究了面向农田物联网复杂事件处理的时空事件模型并定义了合适的事件模型描述语言,首先建立了描述复杂事件模式的时空事件模型,针对多个复杂事件间的关系,通过有向图建立了复杂事件层次模型;针对一个复杂事件内部关系,通过对农田物联网事件常见时序逻辑关系和空间拓扑关系的分析,定义了9类时间关系算符、8类空间关系算符用于判断子事件与复杂事件间的时空关系,并定义了5类时间耦合算符、7类空间耦合算符用于计算复杂事件的时空属性,为描述事件组合模式提供了基础;在此基础上,考虑可读性和易解析性,设计了基于XML的事件模型描述语言;基于案例,通过与其他常用的复杂事件模型相比较,说明了本文事件模型和描述语言更加适用于描述农田物联网复杂事件。

In internet of things in farmlands,control commands are usually triggered by complex events which contain many dimensionalities of information. Complex events are detected by combining mass of atomic events directly sensed by sensors. That is a typical procedure of complex event processing( CEP). As a basis of CEP,a model of complex events,which describes how atomic events construct complex events,is necessary. Current complex event models focus on temporal logic,but spatial logic relations among farmland events are not considered. In this paper,a novel model that considers both temporal logic and spatial logic was proposed and a relevant XML-based specification language was designed. In the model,multi-level complex events were modeled to a directed graph. For one complex event in the graph,based on analysis of temporal and spatial logic relations among component events,nine time relation operators and eight spatial relation operators were defined to describe above relations;five time combination operators and seven spatial combination operators were defined to calculate attributes of complex events from attributes of atomic events. Based on the model,a XML-based language was defined,which balances readability of users,descriptive abilities and difficulties of parsing and compiling. The proposed event model and language were compared to those used in popular CEP systems e. g. SASE +,Esper,etc. From the comparison,the proposed model and language were more suitable for describing complex events in internet of things in farmlands.