基于关联图模型的信息物理融合系统感知数据可信性分析  

Creditability Analysis of Sensor Data in the Cyber-physical System Based on the Relationship Diagram Model

作　　者：汤巍[1] 景博[1] 黄以锋[1] 

机构地区：[1]空军工程大学航空航天工程学院,西安710038

出　　处：《电子与信息学报》2015年第3期679-685,共7页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(51201182);航空科学基金(20101996012)资助课题

摘　　要：针对信息物理融合系统(CPS)感知层数据的不确定性与随机性,该文提出一种CPS中感知数据的可信性分析框架。摒弃以往以传感器为中心的建模思路,该文充分考虑被监测对象因素,建立传感器-目标关联图模型,以此为基础设计了传感数据可信性推理算法。同时,为提高算法的实时性,减少传感器-目标关联图的搜索空间与时间,设计了基于可信目标筛选的改进推理算法。通过实例验证表明,该算法能实时、有效地滤掉CPS中感知数据中的虚假信息,极大提高感知数据的可信性。The high uncertainty and randomness are the characteristics of the sensor data in the Cyber-Physical Systems （CPS）, which make the data unreliable. A creditability analysis framework is proposed to solve those problems. Abandoning the idea that the sensor is the center in modeling, the theory takes monitoring targets into consideration and constructs the sensor-target relationship diagram, which is the base of the creditability reasoning algorithm. Meanwhile, in order to reduce the space and time of searching the relationship diagram, an improving reasoning method basing on filtering the incredible targets is designed. The examples demonstrate that the proposed algorithm can filter out the false message in the sensor data and enhances the creditability of the data in CPS.

关 键 词：信息物理融合系统 感知数据 可信性分析 关联图 

分 类 号：TP301[自动化与计算机技术—计算机系统结构] TP212[自动化与计算机技术—计算机科学与技术]