We investigate declarative process specifications over event data containing both discrete events and continuous data streams (e.g., from IoT sensors). For example, it might be desirable to check LTL-like properties not only over event traces but also over sensor data streams, e.g., a CO \(_2\) value not staying over a certain threshold for some period. As existing LTL approaches are based on discrete state systems, they are not suitable for this hybrid/continuous setting. Therefore, we propose using Signal Time Logic (STL) for verifying temporal properties over a mix of discrete and continuous behaviors. We show how important topics from conformance checking can be expressed in terms of results from signaling and thus be used to check temporal constraints over discrete/continuous systems. In this context, we also introduce a catalogue of Declare templates over STL, as counterpart to the traditional Declare but for continuous settings.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Declarative Process Specifications over Discrete/Continuous Event Data

  • Carl Corea,
  • Anti Alman,
  • Fabrizio Maria Maggi,
  • Paul Hermann Wittlinger

摘要

We investigate declarative process specifications over event data containing both discrete events and continuous data streams (e.g., from IoT sensors). For example, it might be desirable to check LTL-like properties not only over event traces but also over sensor data streams, e.g., a CO \(_2\) value not staying over a certain threshold for some period. As existing LTL approaches are based on discrete state systems, they are not suitable for this hybrid/continuous setting. Therefore, we propose using Signal Time Logic (STL) for verifying temporal properties over a mix of discrete and continuous behaviors. We show how important topics from conformance checking can be expressed in terms of results from signaling and thus be used to check temporal constraints over discrete/continuous systems. In this context, we also introduce a catalogue of Declare templates over STL, as counterpart to the traditional Declare but for continuous settings.