标题： 一种用于从物联网传感器流中检测过程活动的领域特定语言和架构 显示英文标题

标题： A Domain-specific Language and Architecture for Detecting Process Activities from Sensor Streams in IoT

摘要： 现代物联网（IoT）系统配备了大量传感器，提供有关其组件当前操作的实时数据，这对系统的内部控制系统和流程至关重要。 然而，这些数据通常过于细粒度，难以得出物联网系统可能参与的更大流程执行的有用见解。 流程挖掘已开发出用于分析业务流程的高级方法，这些方法也可用于物联网环境中。 将流程挖掘引入物联网需要一个事件抽象步骤，以将低级传感器数据提升到业务流程级别。 在这项工作中，我们的目标是使领域专家能够使用一种新开发的领域特定语言（DSL）称为Radiant来执行此步骤。 Radiant支持在传感器数据中指定指示高层流程活动执行的模式。 这些模式被转换为复杂事件处理（CEP）应用程序，用于在运行时检测活动执行。 我们提出了一种相应的软件架构，用于使用CEP应用程序从物联网传感器流中进行在线事件抽象。 我们评估这些应用程序，以使用物联网传感器在智能制造和智能医疗中监控活动执行。 评估方法和结果向领域专家提供了关于活动检测质量和潜在改进的的信息。 显示英文摘要

摘要： Modern Internet of Things (IoT) systems are equipped with a plethora of sensors providing real-time data about the current operations of their components, which is crucial for the systems' internal control systems and processes. However, these data are often too fine-grained to derive useful insights into the execution of the larger processes an IoT system might be part of. Process mining has developed advanced approaches for the analysis of business processes that may also be used in the context of IoT. Bringing process mining to IoT requires an event abstraction step to lift the low-level sensor data to the business process level. In this work, we aim to empower domain experts to perform this step using a newly developed domain-specific language (DSL) called Radiant. Radiant supports the specification of patterns within the sensor data that indicate the execution of higher level process activities. These patterns are translated to complex event processing (CEP) applications to be used for detecting activity executions at runtime. We propose a corresponding software architecture for online event abstraction from IoT sensor streams using the CEP applications. We evaluate these applications to monitor activity executions using IoT sensors in smart manufacturing and smart healthcare. The evaluation method and results inform the domain expert about the quality of activity detections and potential for improvement.