面向CPS复杂事件流的不确定性研究

信息物理融合系统CPS获得广泛应用需要解决的一个关键问题是软件中的信息处理部分,而复杂事件处理是CPS中信息处理的核心任务之一。CPS环境下的事件具有异构、分散、海量和不确定性等特征。在CPS实际应用中,因噪声、传感器误差、通讯技术等原因而造成的事件不确定性急需解决。为了解决CPS系统中存在的海量不确定事件流问题,提出一种处理不确定事件流的复杂事件处理方法USCEP,该方法不仅可以实时有效地处理海量不确定事件流,还可以有效计算复杂事件的概率。USCEP对现有RFID复杂事件监测方法 RCEDA进行了改进,提供了历史概率事件查询处理的支持,提出一种事件概率模型进行概率计算,并通过关联查询表来提高效率。实验表明,在处理不确定事件流时,该方法比传统方法具有更好的性能。

Research on uncertain CEP for CPS

The wide application of Cyber Physical System (CPS) requires solving a crucial issue regarding the information processing of CPS software.Complex Event Processing(CEP) is one of the core tasks for CPS information processing.The event streams of CPS applications have many features,such as heterogeneity,distribution,mass,uncertainty,and et al. In the practical applications of the CPS,due to reasons,such as noise,sensor error or wireless communication techniques,uncertain complex events processing becomes very important. In order to deal with large scale probabilistic streams,we propose an effective method called Uncertain Complex Event Stream Processing (UCESP),which can not only handle massive uncertain event streams in time but also calculate the probability of complex events.USCESP improves the existing RFID Complex Event Detection Algorithm (RECDA),and supports query processing of history probabilistic events.We also propose an event probability model for probability calculation,and improve the efficiency by using the association query table.Experiments show that, compared with traditional methods, the proposal has better performance in handling uncertain event streams.