Over the last 200 years or so, increasingly sophisticated automation has driven multiple waves of the industrial revolution and lifted the global standard of living by making manufactured goods relatively inexpensive. The next wave, which some call the fourth industrial revolution, includes AI combined with robotics to create intelligent machines capable of learning how to do their job better and more efficiently.
Here at SensiML, we’re doing our part to help enable this 4.0 version of the industrial revolution. Towards that end, we’ve partnered with onsemi to provide AI and machine learning capabilities for their RSL10 Bluetooth 5 radio SoC. This device is already quite capable on its own, with an integrated 32-bit ARM Cortex M3 processor, a 32-bit dual-Harvard DSP core, multi-protocol wireless that includesBluetooth Low Energy, integrated power management, flash and RAM, and a host of peripherals and interfaces.
The RSL10-SENSE-GEVK evaluation kit, including a multi-sensor board that provides developers with a rapid prototyping solution for building ultra-low-power IoT sensing applications around the SoC. The combination of the device and the evaluation kit makes a well-rounded platform for implementing a wide range of industrial IoT applications.is complemented by the
This is where SensiML comes in. Adding our SensiML Analytics Toolkit enables developers to not only build a low-power connected industrial IoT node but also integrate machine learning edge insight transforming connected sensor nodes into smart, learning endpoints. By performing this activity at the edge, reams of raw sensor data can be distilled into nuggets of vital insight for immediate local action as well as efficient low-power wireless transmission cloud analytics and BI dashboard platforms.
Pattern recognition of highly dynamic audio, motion, vibration, current, and voltage sensor data offers a great deal of additional operational insight to industrial processes. Processing this data efficiently right where it is generated allows for practical use across hundreds to thousands of sensor points and also enables real-time local decision making, and greater reliability than centralized cloud analytics. The onsemi RSL10-SENSE-GEVK eval board demonstrates the possibilities with inclusion of light, temperature, pressure, and sound sensors, as well as and gyroscopic sensors. Data from these sources can be used to define optimal versus sub-optimal operating conditions for robotics, factory lines, and manufacturing process equipment to name just a few. Potential maintenance issues can be identified ahead of system failure and actively managed to avoid catastrophic equipment failure and a resulting “lines down” situation.
While there are other ML ops workflows meant for real-world production projects. Our solution includes neural networks along with classic machine learning algorithms to provide the most compact models while still meeting the needs of the application. Thus, we are the perfect fit for edge IoT in industrial automation environments – providing high AI “IQ” in a small space with extremely low power consumption. A great complement for onsemi’s RSL10 and associated sensor development kit and our way of helping to drive the next wave of the industrial revolution.solutions for the edge, SensiML combines industry-leading code optimization with thorough
To learn mode about the SensiML and onsemi solution for smart Industrial edge sensor nodes, register and view the hands-on demo here:
a joint webinar presentation by onsemi and SensiML.