Infineon Pushes Machine Learning to the Edge With New MCU Family

As edge devices evolve beyond simple tasks, they struggle to access the computing power necessary to support smart homes, IoT, or wearable devices. Infineon has announced the PSoC Edge family of microcontrollers to provide high performance and security for devices operating machine learning (ML) at the edge.

 

The PSoC Edge family of MCUs offers built-in ML support, Autonomous Analog, and many more features that make it a candidate for next-generation smart edge devices. Image used courtesy of Infineon

 

This article takes a closer look at the features of the PSoC Edge family to learn how it can be used to improve the performance of edge devices and improve ML capabilities for novel applications.

 

Enhanced ML Performance

A key spec of the new PSoC Edge family is the onboard ML hardware acceleration. In addition to Cortex-M55-enabled DSP support and an Ethos-U55 NPU, the PSoC Edge devices include Infineon’s proprietary neural network hardware accelerator NNLite. This hardware acceleration allows the family to be used in a variety of applications requiring responsive computing and control. Furthermore, a new "Autonomous Analog" feature allows designers to perform versatile always-on sensing and audio processing for IoT applications.

 

The Autonomous Analog feature of the PSoC Edge family allows designers to constantly monitor analog inputs without requiring high power consumption. Image used courtesy of Infineon

 

Interested developers can leverage the ModusToolbox and Imagimob Studio Edge AI software to simplify the development process, including end-to-end ML development. The processing power is complemented by on-chip memory, including non-volatile RRAM and secure, high-speed external memory support. Furthermore, Infineon designed PSoC Edge with scalability in mind, allowing designers to reuse software to reduce time-to-market and development costs.

 

Powerful and Secure Applications

The PSoC Edge family offers embedded security options through Infineon’s Edge Protect technology, which protects data produced or stored onboard. 

 

PSoC Edge domain system architecture. Image used courtesy of Infineon

 

Infineon has identified several applications that would benefit from the PSoC Edge family, including smart home devices, wearables, and human-machine interaction designs. Outside of these specific areas, however, designers could make use of the PSoC Edge MCUs in any application that requires on-chip ML acceleration. In the case of an IoT device, ML features could be used to adaptively interact with the environment, while robotics use cases could leverage the ML capabilities to improve efficiency and performance for autonomous devices.

 

Bringing ML to the Edge

As trends continue toward allocating more computing power to edge devices versus centralized servers, new devices such as the PSoC Edge family lend more power and security to personal devices. 

The PSoC Edge family is currently available to early-access customers. As more developers use the chip, they can evaluate the performance increases offered by the on-chip ML acceleration and Autonomous Analog signal chains.