Delivering ultra-low-power industrial edge AI sensing applications

SensiML Corporation has teamed with onsemi to produce a complete machine learning solution for autonomous sensor data processing and predictive modelling. The collaboration combines its Analytics Toolkit development software with the RSL10 Sensor Development Kit from onsemi to produce a platform perfect for edge sensing applications, including industrial process control and monitoring. Its ability to support AI functions in a small memory footprint, together with the advanced sensing and BLE connectivity offered by the RSL10 platform, with no need for cloud analytics of highly dynamic raw sensor data.

Providing the industry’s lowest power BLE connectivity, the kit combines the RSL10 radio with a full spectrum of environmental and inertial motion sensors onto a tiny form-factor board that interfaces easily to the Toolkit. Developers utilising the RSL10-based platform and the SensiML software together can simply add low latency local AI predictive algorithms to industrial wearables, process control, robotics, or predictive maintenance applications despite their expertise in data science and AI. The resulting auto-generated code allows smart sensing embedded endpoints that transform raw sensor data to critical insight events directly where they occur and can take relevant action in real-time. Moreover, the smart endpoints also drastically lessen network traffic by communicating data only when it provides valuable insight.

“Cloud-based analytics is too slow, too remote and too unreliable for the most critical industrial processes, said Dave Priscak, vice president of Applications Engineering at onsemi. “The difference between analysing a key event with local machine learning versus remote cloud learning can equate to production staying online, equipment not incurring expensive failures and personnel staying safe and productive.”

“Other AutoML solutions for the edge rely only on neural network classification models with only rudimentary AutoML provisions, yielding suboptimal code for a given application,” said Chris Rogers, SensiML’s CEO. “Our comprehensive AutoML model search includes not only neural networks, but also an array of classic machine learning algorithms, as well as segmenters, feature selection, and digital signal conditioning transforms to provide the most compact model to meet an application’s performance need.”