Enabling sensor data capture and ML for smart home fitness and industry applications

10-02-2022 | Infineon | New Technologies

Infineon Technologies AG collaborates with SensiML to offer developers an easy and seamless process to capture data from Infineon XENSIV sensors, train ML models, and deploy real-time inferencing models directly on ultra-low-power PSoC 6 MCUs. This can be achieved by using the SensiML Analytics Toolkit and ModusToolbox. This collaboration presents designers with the right tools to develop smart applications for IoT devices spanning the smart home, industrial and fitness sectors.

Infineon also hosted the "Build AI for the IoT Design Challenge," launched on hackster.io on 3 February 2022, where innovative developers were challenged to employ a combination of tools to develop new ML/AI solutions. Design challenge participants used Infineon's ModusToolbox ML, ultra-low-power PSoC 6 technology and CAPSENSE capacitive sensing integrated with the robust connectivity of the AIROC wireless connectivity solutions and sensors in the XENSIV family. SensiML software offered participants the required firmware and data science tools to capture and label data from sensors. Additionally, they used an AutoML cloud platform to train models without extensive AI expertise and then deploy the resulting models on the PSoC 6 MCU for use on edge devices.

"Capturing the right data for smart home, fitness and industrial IoT devices, and implementing this information to train ML modules is a complex but key process for OEMs to make their devices smarter," said Steve Tateosian, vice president of IoT Compute and Wireless Business Line of Infineon. "Our collaboration with SensiML enables just that by combining their ML/AI software tools with our proven PSoC 6 MCUs, connectivity and sensor solutions to simplify the complexity of creating intelligent IoT solutions. Through this collaboration, we look forward to offering a complete suite of the right tools to enable developers and design challenge participants to create tomorrow's smart IoT devices for a variety of sectors."

"Infineon offers a broad range of low-power processors and cutting-edge sensors beneficial to developers across a vast array of embedded IoT applications," said Chris Rogers, CEO of SensiML. "Our SensiML Data Analytics toolkit can help customers using those Infineon products to quickly and easily implement AI/ML capabilities to transform sensor-enabled devices into smart IoT solutions."

Ideal for battery-powered applications, its PSoC 6 MCUs are constructed on an ultra-low-power architecture offering low-power design techniques. The dual-core Arm Cortex-M4 and Cortex-M0+ architecture enable designers to optimise power and performance simultaneously. The PSoC 6 MCU delivers the highest level of protection defined by the PSA from Arm and utilises its dual cores combined with configurable memory and peripheral protection units. With SensiML's AI software tools, customers can now capture the correct sensor data and create ML models to run on the PSoC 6 MCU.

By Natasha Shek