AI is everywhere, and with the rise of machine learning and the Big Data revolution, it is a transformational field essential for building a smart and connected world.
STMicroelectronics is a prominent player in the semiconductor industry, offering a diverse range of solutions for various applications, including industrial, automotive, consumer and personal electronics. They are highly known for their comprehensive offerings including microcontrollers, sensors, analog chips, and embedded systems, which are integral to the development of AI-enabled devices operating at the edge.
Matteo MARAVITA, Senior Manager, AI Competence Center and Smartphone Competence Center, Asia Pacific Region, STMicroelectronics, presented before the media, their latest release – ST Edge AI Suite. The article is an excerpt from the online briefing.
In today’s interconnected world, edge AI is becoming increasingly essential for businesses looking to upgrade their products with intelligence and decision-making capabilities. By processing data locally at the edge, companies can reduce latency, improve privacy, and enhance overall efficiency. ST with its extensive portfolio of hardware solutions along with the capabilities of Edge AI Suite, is well positioned to facilitate this transition to a more intelligent edge. Companies can unlock a wide range of possibilities including real-time decision-making, customisable solutions pertaining to industry needs, scalability and future-proofing etc.
ST’s Automotive Edge AI Solutions
ST has partnered with HPE Group to create a virtual sensor that optimizes the operation and maintenance of EV motors. To explain, the AI algorithm works on their latest automotive controller, the Stellar family, which takes external data from sensors and uses it to extrapolate and estimate the internal rotor temp of the motor. On the side, the microcontroller not only runs several AI algorithms (also for predictive maintenance to identify anomalies) but also drives the motor itself.
ST’s Consumer Edge AI Solutions
Using AI algorithms and STM32 microcontroller at the helm of the solution, ST’s innovation has helped businesses experience 15-40% performance improvement for washing machines. To explain, two machine learning algorithms work towards creating a virtual sensor approach and in collecting data from a 6-axis motion sensor to enable drum collision avoidance. Through the algorithm input, the motor is driven using exactly the needed current, and the water and detergent requirement is adjusted, thus saving significant energy and water for a washing cycle. Both algorithms have been developed with NanoEdge AI and run on an STM32G0 MCU together with an ST 6-axis sensor.
ST’s Personal Electronics Edge AI Solutions
The HP engineering team collaborated with ST in developing and training AI models that recognize different user activities based on device and user motion. Different scenarios were addressed, including – the laptop is placed on a table, on the user’s lap, carried inside a bag, and taken out. The team worked on Smart context detection using smart sensor technology to optimize the power monitoring of the laptop by avoiding overheating and battery drain. All solutions for ultra-low power PC activity monitoring are based on ST’s 6-axis IMU MEMS sensors.
Implementation on edge AI- Tackling Software and Hardware Challenges
For developers to create a full-fledged edge AI solution, ST has put in years of R&D efforts to discover and evaluate various bottlenecks on both the hardware and software side of development. This included research on machine learning techniques, development of ML models, and testing the performance, security, and power efficiency of the connected devices with overall integration of the system.
Furthermore, ST Edge AI Suite addresses the needs and requirements of different profiles. It empowers embedded developers, data scientists, and product designers and creators with various aspects of ML model optimization and product redefining.
ST has announced the ST Edge AI Suite, a comprehensive and integrated set of software tools free to use with ST hardware. This offers developers and companies an ecosystem with a broad range of hardware with free software and tools, supported by partnerships with cloud services and AI toolchain providers. The ST Edge AI Suite is set to simplify the development of AI solutions exploiting ST’s range of hardware and related tools for embedded AI optimizations.
The ST Edge AI Suite is compatible with the external ecosystems for AI development, simulation tools like MATLAB, machine learning models trained using deep learning frameworks like TensorFlow Lite, Keras, PyTorch, etc., and a possibility of connection to cloud services like AWS and Azure.
Also, the tool works across multiple ST hardware platforms including STM32 general-purpose MCUs, STM32N6 and STM32 MPUs built for industrial applications, Stellar automotive microcontrollers, etc.
ST’s strategy on AI relies keenly on an innovative, unified optimizer of embedded AI solutions called ST Edge AI Core Technology. Looking at the ST Edge AI Core is a critical component that brings together all the software and tools engineers need at each step of their project. It is the core library with a unified common line interface that the customer could use to evaluate the model and further port it to the specific target device. The software tools can be used with the STM32 microcontroller and the MEMS sensors with the hardware accelerators (MLC or ISPU).
Also, the tools (ST Edge AI & Nano AI Studio) are completely free for unlimited quantities on any STM32.
