Abstract: In response to the challenging task of vital information perception and collection in the thriving realms of physical and cyber spaces,a unified method for constructing virtual sensor models was proposed.By scrutinizing the structural attributes of tangible sensors and employing ontological methods for construction,a unified framework comprising ontological characteristics,transformation modules and relational properties had been distilled.Based on these definitions and the requisites of application processes,six guiding principles were established,culminating in three distinct construction approaches.These methods expedited the rapid development of virtual sensors to meet diverse requirements,thereby enhancing the operability of models and broadening the scope of information gathering.This construction approach expanded the applicability of virtual sensors.The model was implemented in the production environment of a medium-voltage cable production line that necessitated the collection of a myriad of data types and large volumes of information,thus various forms of information including geometric,physical,behavioral and rule-based data were successfully extracted,and the spectrum of data perception and collection within production context of the cable factory was effectively amplified.

Key words: virtual sensor, ontology model, modeling approach, construction criteria, data acquisition

摘要： 针对蓬勃发展的物理和赛博空间中出现的关键信息感知采集困难的问题,提出了一种统一的虚拟传感器模型构建方法。通过分析实体传感器的结构特点,并利用本体论法进行构造,提炼出包括本体特征、转化模块和关联特性三大要素组成的统一结构。基于这些定义和应用过程的需求,建立六大构建准则,归纳整理出的3种不同的构建方式,实现了对多种需求的虚拟传感器的快速构建,使得模型的可操作性得到增强,扩大了信息采集的范围。这种构建方式扩大了虚拟传感器的适用范围,本文将该模型应用于需采集众多种类和大量数据的电缆厂中压生产线生产环境中,成功得到各种形态的信息,包括:几何、物理信息以及行为、规则信息,有效扩大了数据感知采集的范围。

关键词: 虚拟传感器, 本体模型, 建模方法, 构建准则, 数据采集