Silicon Labs uses Matter to deliver AI and machine learning to the edge
The complete Pro Kit for the new BG24 and MG24 SoCs with all the necessary hardware and software for developing high-volume, scalable 2.4 GHz wireless IoT solutions. The new hardware supports Matter, ZigBee, OpenThread, Bluetooth Low Energy, Bluetooth mesh, proprietary and multi-protocol operation. (Photo credit: Silicon Labs)
//

Silicon Labs uses Matter to deliver AI and machine learning to the edge

3 mins read

Texas-based electronics company Silicon Labs is bringing AI and machine learning to the edge with a platform that uses the Matter smart-home connectivity standard.

The BG24 and MG24 2.4GHz wireless systems-on-chip (SoCs) are for Bluetooth and multiple-protocol operations, respectively. This hardware and software architecture could help bring AI and machine learning applications to battery-powered edge devices, as well as wireless high performance.

The low-power families are matter-ready, supporting several wireless protocols and including PSA level-three Secure Vault protection, making them ideal for a wide range of smart home, medical, and industrial applications.

The SoCs have Matter, Zigbee, OpenThread, Bluetooth Low Energy, Bluetooth mesh, proprietary, and multi-protocol support, as well as embedded AI and ML accelerators.

Advertisement

A software toolkit is also available to assist developers in fast developing and deploying AI and machine learning algorithms. It makes use of some of the most popular toolkits, including TensorFlow.

“The BG24 and MG24 wireless SoCs represent an awesome combination of industry capabilities including broad wireless multiprotocol support, battery life, machine learning and security for IoT edge applications,” said Matt Johnson, CEO of Silicon Labs.

AI and machine learning have the potential to provide even more intelligence to edge applications such as home security systems, wearable medical monitors, sensors monitoring commercial sites, and industrial equipment, according to IoT product designers. However, organizations considering deploying AI or machine learning at the edge face significant performance and energy consumption penalties that may outweigh the benefits.

Advertisement

The BG24 and MG24 are designed to reduce the severity of these penalties. Internal testing revealed a four-fold increase in performance and a six-fold increase in energy efficiency thanks to the hardware, which is intended to handle complex computations swiftly and effectively. Because the machine learning calculations are performed locally rather than in the cloud, network latency is reduced, allowing for faster decision-making and action.

The families also have a lot of flash and ram. This means they can evolve to handle several protocols, Matter, and trained machine learning algorithms for big datasets. Secure Vault, which is PSA level three certified and is the highest level of security certification for IoT devices, provides the protection needed in products like door locks, medical equipment, and other sensitive deployments that were hardening the device against external attacks is critical.

In addition to natively supporting TensorFlow, Silicon Labs has partnered with AI and machine-learning tool providers such as SensiML and Edge Impulse to provide developers with an end-to-end toolchain that simplifies the development of machine-learning models optimized for embedded wireless application deployments.

Advertisement

Developers may use this AI and ML toolchain in collaboration with Silicon Labs’ Simplicity Studio and the BG24 and MG24 SoCs to create apps that pull data from a variety of connected devices and communicate with one another via Matter to make intelligent machine-learning-driven decisions.

Many lights in a commercial office building, for example, are controlled by motion detectors that decide whether the lights should be turned on or off based on occupancy. Workers may be left in the dark when typing at a workstation with only hand and finger activity since motion sensors alone are unable to detect their presence.

The additional audio data, such as the sound of typing, can be passed through a machine-learning algorithm to allow the lighting system to make a more informed choice about whether the lights should be on or off, by linking audio sensors with motion detectors through the Matter application layer.

Sensor-data processing for anomaly detection, predictive maintenance, audio pattern recognition for improved glass-break detection, simple-command word recognition, and vision use cases such as presence detection or people counting with low-resolution cameras are all enabled by ML computing at the edge.

Advertisement

In a closed alpha program, more than 40 organizations from various industries and applications have already begun creating and testing this platform.

With more accurate asset tracking, real-time price updating, and other uses, global retailers are working hard to improve the in-store shopping experience. Commercial building management participants are looking at how to make their building systems, such as lighting and HVAC, more intelligent in order to save money and lessen their environmental imprint. Finally, consumer and smart home suppliers are aiming to make it easier to connect various devices and broaden the ways in which they interact in order to give consumers new features and services.

The single-die BG24 and MG24 SoCs include a 78MHz Arm Cortex-M33 processor, 2.4GHz radio, 20bit ADC, a combination of flash up to 1536kbyte and RAM up to 256kbyte, and an AI and ML hardware accelerator for processing machine-learning algorithms while offloading the Cortex-M33, allowing applications to focus on other tasks.

Advertisement

The SoCs are delivering today to alpha users in 5 by 5mm QFN40 and 6 by 6mm QFN48 packages and will be available for mass deployment in April 2022. Designers working on applications have access to a variety of evaluation boards. In the second half of 2022, modules based on the SoCs will be available.

Leave a Reply