Method to Identify Process Activities by Visualizing Sensor Events.

With the onset of the Internet of Things (IoT) everyday objects suddenly become data sources equipped with sensors measuring the object’s properties and surroundings. However, the lack of process-awareness in IoT environments (e.g., smart factories) prevents the adoption of more sophisticated process analysis and optimization. One hurdle is the differing abstraction level of low-level sensor data and process-level activities. We propose a method to identify activities step-by-step from raw IoT data using visualizations. By relying on minimal process knowledge, we discover process activities from sensor events. These activities are represented by specific sequences of sensor events–Activity Signatures–that serve as a basis for finding similar activities. We demonstrate the method’s validity with a proof of concept in a smart factory.