TL;DR:
- TDK unveils InvenSense SmartEdgeML, a pioneering machine learning tool for a 6-axis IMU.
- SmartEdgeML empowers developers to add machine learning models to wearables, hearables, AR glasses, IoT, etc.
- It operates at the IMU sensor chip level and consumes less than 30 µA of power.
- The launch took place at CES 2024 in Las Vegas.
- The innovation reduces raw data processing, enhancing device battery life, data privacy, and system latency.
- TDK also introduces the InvenSense SmartBug 2.0, aiding users in evaluating the SmartEdgeML ecosystem.
- Three core components: SIF (sensor inference framework), ICM-45686-S IMU, and MD-45686-ML multi-sensor wireless module.
- The ICM-45686-S IMU offers premium temperature stability and vibration rejection.
- Ideal for applications like AR glasses, VR, OIS, drones, TWS, and robotics.
Main AI News:
TDK has unveiled a groundbreaking machine learning development tool, known as the InvenSense SmartEdgeML, tailored for a revolutionary 6-axis inertial measurement unit (IMU) device and module. This innovative tool empowers developers to seamlessly integrate decision tree machine learning models into a wide array of products, including wearables, hearables, AR glasses, and IoT devices, all at the IMU sensor chip level. Notably, it stands out as the first solution capable of generating and executing machine learning models within the compact confines of a 2.5 x 3mm 6-axis motion sensor IMU, all while consuming less than 30 µA of power.
This groundbreaking ML development tool, alongside the new IMU sensor and module, has been officially unveiled on the grand stage of CES 2024 in Las Vegas. As Sahil Choudhary, Director of Motion Sensors and Software at InvenSense, aptly puts it, “TDK’s SmartEdgeML is a paradigm shift in edge machine learning, as it will allow developers, ODMs, and OEMs to implement ML-optimized motion sensor algorithms on an IMU sensor chip. This reduces the amount of raw data going to edge processors, which significantly improves device battery life, data privacy, and system latency.”
TDK has also introduced the InvenSense SmartBug 2.0 (MD-45686-ML), a versatile multi-sensor wireless module that includes the InvenSense ICM-45686-S IMU. This module serves as an ideal evaluation platform, enabling users to initiate their journey with the InvenSense Sensor Inference Framework (SIF) and the ICM-45686-S IMU. While the SIF is readily available for download, the MD-45686-ML and ICM-45686-S are set to hit distributors’ shelves by February 1, 2024.
The SmartEdgeML ecosystem comprises three essential components, seamlessly integrated for a holistic experience:
- Sensor Inference Framework (SIF): This software component serves as a comprehensive hub, offering users a one-stop solution for collecting IMU sensor data, selecting customized features, building ML models, testing ML performance, deploying, and executing those models on the ICM-45686-S IMU via the SmartBug 2.0. Tested applications encompass algorithms like exercise classification (such as squats, jumping jacks, lateral raises, or push-ups) and wrist gesture classification (including fight, turn, shake, or still).
- ICM-45686-S IMU: The hardware component of this ecosystem, this diminutive 2.5 x 3mm IMU from the TDK BalancedGyro family, boasts the remarkable capability to run ML decision tree models on-chip, all while maintaining a power consumption of under 30 µA. Its exceptional temperature stability and vibration rejection make it the ideal choice for applications spanning AR glasses, VR, OIS, drones, TWS, and robotics, where the demand for high-performance and ultra-low power machine learning algorithms is paramount.
- MD-45686-ML: This all-inclusive multi-sensor wireless module includes the ICM-45686-S 6-axis motion sensor and is fully compatible with the SIF. Its compact form factor and versatile BLE + USB interface, embodied in the SmartBug 2.0, facilitate a rapid transition from data collection to ML model development, ultimately culminating in seamless deployment on the ICM-45686-S IMU. For those embarking on the SmartEdgeML journey, this device stands as the quintessential choice, expediting the path to innovation.
Conclusion:
TDK’s SmartEdgeML, along with the 6-axis IMU technology, represents a transformative leap in edge machine learning. It not only empowers developers but also enhances device performance, privacy, and power efficiency. This innovation is poised to reshape the market, particularly in wearable tech, AR, IoT, and beyond, by opening new possibilities for seamless integration of machine learning in compact devices.
