Cartesiam announces AI tools to classify machine anomalies, not only detect problems

With tools from Cartesiam, developers of embedded devices can equip Arm Cortex-M microcontrollers to detect and classify anomalies in a wide variety of industrial and medical machines. (Cartesiam)

 

Edge AI tools are already being used to predict when a machine or device is about to fail. Now the technology is advancing to saying more specifically what is wrong, not only that there’s a generic problem.

The latest tools can help move from saying “your dish washer isn’t draining properly” to “your drainage problem is because a belt is wearing out.”

Free Daily Newsletter

Interesting read? Subscribe to FierceElectronics!

The electronics industry remains in flux as constant innovation fuels market trends. FierceElectronics subscribers rely on our suite of newsletters as their must-read source for the latest news, developments and predictions impacting their world. Sign up today to get electronics news and updates delivered to your inbox and read on the go.

Paris-based Cartesiam announced on Tuesday a means to classify such anomalies with software on embedded devices built with Arm Cortex-M microcontrollers.

The company launched NanoEdge AI Studio V2 that builds on its first version, promising it can take AI systems from detecting an anomaly in its first generation to recognizing what the anomaly is. Pricing was not announced.

In one example, Cartesiam CEO Joel Rubino said in an interview that version 1 of Studio would be able to tell a plant manager that a pump was about to break down, while version 2 would be able to add that the pump is about to break down due to a failure on a ball bearing.

Such insight can run on a Windows or Linux PC without a connection to the cloud, he said. “It works like a search engine, taking all the different elements and combining them for the best information,” he added. “Our unique value proposition is we don’t come from the cloud world with a big framework but are designed with the cloud in mind working with limited power and memory and a huge data set.

Naval Group, an industrial contractor and integrator of warships and combat systems, has deployed Cartesiam tools precisely because it didn’t want to send prediction data out to the cloud, he said.

Cartesiam faces competitors in machine learning tools that can provide preventive maintenance “but I don’t see a lot of companies that have effectively done preventive maintenance,” Rubino added.

 The software already runs on thousands of devices with a variety of customers such as Schneider Electric and Lacroix Electronics. NKE Watteco uses Cartesiam technology in Bob Assistant, an intelligent sensor box already installed on dozens of industrial sites in Europe to provide predictive maintenance for pumps, engines and more.

Also on Tuesday, Cartesiam announced a collaboration with Bosch Connected Devices and Solutions GmbH to add NanoEdge AI Studio to its Cross Domain Development Kit, the XDK. The kit has eight sensors to remotely monitor and control processes with Bluetooth or Wi-Fi.

Cartesian also announced Use Case Explorer, a web-based platform to allows users to download real dataset to try out the NanoEdge Ai Studio on various use cases such as medical ventilator obstruction detection and breast cancer detection.

RELATED: AI at the Edge