Anomaly detection is implemented via built-in functionality that can be deployed wherever anomaly detection is required.

Typically, anomaly detection functionality is built into an edge application, either on or near the device being monitored. When the anomaly detection component is deployed, it observes the data signal from a device, learns what the expected data stream looks like, and then monitors for data points that fall outside of the expected range. For example, by monitoring sensor data from a device, like a generator or an engine, anomaly detection can detect changes in the device operation and determine whether those changes signal that the device is in an anomalous state.

Anomaly detection can detect anomalies on high and low data rate signals.

Anomaly detection can be activated through the alert process in ThingWorx Composer. For information about implementing anomaly detection in ThingWorx, see Implementing Anomaly Detection.