Artificial Intelligence (AI) is a big thing in the world of IoT. As Things become smarter, companies are looking more and more to AI to leverage the value of the data being collected, and so we’re excited to announce the launch of Thingstream’s first AI Anomaly Detector node.
Thingstream Anomaly Detector
Not everyone has the luxury of a fully featured IoT platform with built-in AI features which is why we’ve added the new Anomaly Detector node to our Data Flow Manager.
How it works
In this first release, customers can simply drop the node at the start of a flow subscribed to the publish topic of their Things, where it will learn the typical publish frequency of each Thing that calls the flow.
Once 5 data points have been collected for a Thing, a monitor thread will start to check for absence, i.e. a case where the Thing has not published at the normal frequency, allowing a tolerance of 46%. If absence is detected, a message is published to a dedicated alarm topic which another flow can be subscribed to in order to raise alarms by sending an email, SMS or posting an alarm state to the customer’s enterprise systems. When the Thing checks in again, the monitor thread will detect the renewed presence, and send an OK state to the alarm topic for onward communication.
You can find information about how the node operates in the “Info” tab of the Data Flow Manager, and pricing information our detailed pricing page.
What’s coming next?
Over the coming months, we will enhance the AI functionality so that you can choose other dimensions to monitor, and configure your own learning parameters, alarm topics etc. so that you can start to explore and unleash the power of the data being collected by your Things simply, and efficiently; the Thingstream way.
Look out for more exciting announcements, and of course, if you have any questions or would like more information, please get in touch.