Two-Dimensional Rule Language for Querying Sensor Log Data: A Framework and Use Cases

Motivated by two industrial use cases that involve detecting events of interest in (asynchronous) time series from sensors in manufacturing rigs and gas turbines, we design an expressive rule language DslD equipped with interval aggregate functions (such as weighted average over a time interval), Allen’s interval relations and various metric constructs. We demonstrate how to model events in the uses cases in terms of DslD programs. We show that answering DslD queries in our use cases can be reduced to evaluating SQL queries. Our experiments with the use cases, carried out on the Apache Spark system, show that such SQL queries scale well on large real-world datasets.

Authors

• Sebastian Brandt 0001
• Diego Calvanese
• Elem Güzel Kalayci
• Roman Kontchakov
• Benjamin Mörzinger
• Vladislav Ryzhikov
• Guohui Xiao 0001
• Michael Zakharyaschev

Keywords

• Ontology-based data access
• temporal logic
• sensor log data