AI REGIO CUSUM RLS filter service
CUSUM RLS filter contains a change detection algorithm for multiple sensors, using the Recursive Least Squares (RLS) and Cumulative Sum (CUSUM) methods [F. Gustafsson. Adaptive Filtering and Change Detection. John Willey & Sons, LTD 2000].
The asset contains a change detection algorithm for multiple sensors, using the Recursive Least Squares (RLS) and Cumulative Sum (CUSUM) methods
As an AI resource the “CUSUMRLSfilter" asset is currently implemented as Open Source Solution whose main aim is to detect abrupt changes on the measurements recorded by a set of sensors.The asset was implemented as part of one of the experiment of the AI REGIO project, and subsequently adapted for general use.
The AI REGIO project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreements No 952003 (AI REGIO).
The service has following endpoints:
- GetAlarms: endpoint that returns the alarms
- The endpoint takes in a vector of float values "measured_values" representing the current values readings from the sensors.
- Then applies the CUSUM RLS filter to the input values: For each sensor, it calculates the estimated value using RLS method and the error between the estimated value and the real value. It then applies the CUSUM algorithm to the error to determine whether an abrupt change is detected in any of the sensors.
- The endpoint also keeps track of the number of consecutive abrupt changes on each sensor and triggers an alarm if the maximum number of consecutive changes is exceeded.
- The output of the endpoint is a vector of boolean values “activated_alarms” that indicates whether an alarm is triggered by a sensor.
- Config: endpoint to configurate the service
- GetCredit: endpoint for project credits
This service, implemented and accessible through the AIoD Experiment Platform, can be used by any device that monitors different sensors even if it has a limited amount of memory and processing capacity.