▲ Moon Hyun-soo, ST Korea GPM Division Technology Manager (Photo courtesy of ST)


X-CUBE-MATTER solution, CSA compatibility tested and continuously updated
Ecosystem and AI solution support to accelerate implementation and development of edge AI processing

STMicroelectronics (hereinafter referred to as ST) introduced software packages that support ST's products and ST MCU and MPU-based edge AI (on-device AI) solutions.

ST held a press conference at the Novotel Ambassador in Gangnam on the 19th and explained the new STM32 products and solutions to be released today.

At this event, Kamel KHOLTI, Wireless Product Line Manager at ST, and Junho Lee, Marketing Manager of GPM Wireless Division at ST Korea, introduced the new software package, X-CUBE-MATTER, and Hyunsoo Moon, Technology Manager at GPM Division at ST Korea, explained ST MCU and MPU-based edge AI solutions.

The X-CUBE-MATTER solution is a software package that provides easy integration of the Matter SDK (Software Development Kit), which is released two to three times a year, to the STM32 family that supports Matter technologies such as Thread, Wi-Fi, and Ethernet.

Camel Colty Manager and Junho Lee Vice President said that ST is developing network standards in the Connectivity Standard Alliance (CSA).He explained that the organization is registered and actively operating.

They said that ST's SDK is continuously updated through various tests with ecosystem participants in this group, and that it is very excellent in terms of compatibility and stability.

The SDK also includes Matter features such as OTA, provisioning services, and low-power endpoints, and supports Linux, Mac, and Windows as development environments, providing broad accessibility.

A development kit fully compatible with the STM32 ecosystem tools is included to enable easier prototyping and development.

A wide range of application examples are also available, including end devices such as lighting, window coverings, thread border routers, and Matter to non-Matter bridges.

Manager Moon Hyun-soo explained the STM32 AI solution, saying that ST supports an ecosystem and various AI solutions that accelerate the implementation and development of edge AI processing based on STM32 MCUs and MPUs.

Processing pre-trained AI models using ST's MCU and the computing performance of the MPU itself has various advantages.

It has the advantage of implementing AI processing directly on STM32 MCU and MPU devices without connecting to the cloud or host PC, ensuring real-time performance, low latency, and improving security performance because data is processed internally.

ST has been investing and developing edge AI processing for about 10 years, culminating in the STM32Cube.AI platform in 2018.

This platform converts AI models running on the cloud and hosts based on STM32 MCUs into C code format so that they can run on STM32 MCUs.It is still being continuously developed as a tool.

NanoEdge.AI Studio, released about two years ago, is an automated autoML program that automatically creates AI models, C code models, etc. by providing only a data set, even if you have no knowledge of AI.

ST provides the above ecosystems free of charge to STM32 users.

In addition, ST is distributing AI models in advance through ST Model ZOO, such as Vision AI, such as Image Classification and Object Detection, and AI models using data from Audio and various sensors, as pre-learned AI models, and is supporting users to apply these directly to applications or shorten development time through transfer learning of already created AI models.

Manager Moon Hyun-soo concluded his presentation by saying, “On-device AI is a hot topic in the market these days,” and “The key is to develop lightweight technology that can operate on MCUs and MPUs without loss. ST will continue to support customers based on the know-how accumulated over the past 10 years.”