At last week’s Embedded World in Germany, Japanese electronics company Renesas had live demonstrations of artificial intelligence (AI) and machine learning (ML) on an MCU based on the Arm Cortex-M85 processor.

The demos showed the performance uplift in AI and ML applications made possible by the Cortex-M85 core and Arm’s Helium technology.

At Embedded World last year, Renesas became the first company to demonstrate working silicon based on the Cortex-M85. This year, it showed the features of the processor in AI use cases.

The first demonstration showed a people detection application developed in collaboration with Plumerai, a specialist in vision AI, that identified and tracked persons in the camera frame in varying lighting and environmental conditions. The TinyML models used in this application can lead to low-cost and lower power AI for a wide range of IoT implementations.

“This really gives a big performance increase for DSP,” said Graeme Clark, business unit manager at Renesas. “Plumerai has made a vision system that for the first time does people recognition on an embedded microprocessor. This could be used for security systems. Or it can be used in shops to recognise who is going in and see when they take things off the shelves.”

The second demo showed a motor control predictive maintenance use case with an AI-based unbalanced load detection application using Tensorflow Lite for microcontrollers with CMSIS-NN.

Delivering over six CoreMark/MHz, Cortex-M85 enables IoT use cases that require high compute performance and DSP or ML capability, realised on a single, simple-to-programme Cortex-M processor. The Cortex-M85 processor features Helium technology, Arm’s M-Profile vector extension, available as part of the Arm v8.1M architecture. It delivers a performance uplift for ML and DSP applications, accelerating compute-intensive applications such as endpoint AI.

Both demos showed the performance uplift made possible by the application of this technology in AI use cases.

Renesas will implement the Arm processor within its RA family of MCUs.

The new Cortex-M85 core supports Arm TrustZone technology for protection of secure assets. Combined with TrustZone, Renesas’ integrated cryptographic engine, immutable storage, key management, and tamper protection against DPA and SPA side-channel attacks provides integrated secure element functionality.

Renesas also announced the expansion of its 32bit RA family with two groups based on the Cortex-M33 core with TrustZone technology. The 100MHz RA4E2 and 200MHz RA6E2 groups are optimised to provide better power efficiency without compromising performance.

With 128kbyte and 256kbyte flash options and 40kbyte SRAM, the groups integrate connectivity options such as on-chip Can-FD, USB, QSPI, SSI and I³C interfaces and offer an upgrade path to other members of the RA family. They are suitable for applications requiring high performance in small packages such as sensing, gaming, wearables and appliances.