Infineon – PSOC™ Edge MCUs – PTM | DigiKey

The PSOC Edge E84 AI Kit (KIT_PSE84_AI) is a hardware platform designed for rapid prototyping and development of Machine Learning (ML) and Edge AI applications. It features the PSOC ™ Edge E84 series microcontroller (MCU) and a multitude of on -board multimedia, Machine Learning (ML), and connectivity features. It also supports Infineon's DEEPCRAFT™ AI Software Suite and ready-to-deploy ML models.

PSOC Edge devices use a heterogeneous multi-core architecture split into a high performance domain and a low power domain.

The high-performance domain features an Arm Cortex-M55 at up to 400 megahertz. This includes Arm Helium technology, which provides DSP extensions for improved signal processing performance for use cases such as audio, imaging, power electronics, and communications. It also has an Arm Ethos-U55 neural processing unit and can provide up to 480 times higher ML performance over existing Cortex-M based systems, supporting vibration detection, anomaly detection and keyword, speech, image, object, and gesture recognition. The high-performance domain can be disabled to reduce power consumption.

The low-power domain uses an Arm Cortex-M33 at up to 200 megahertz paired with Infineon’s NNLite NPU. This proprietary ML hardware accelerator is optimized for efficient data handling, ultra-low power consumption, and use with low to moderate complexity ML models, making it suitable for always-on AI applications, such as wake word detection or human activity recognition.

The secure enclave implementation provides secure boot, secure firmware updates, key storage functions and crypto operations. It achieves compliance with Infineon’s Edge Protect Category 4

For HMI, the MCUs integrate a 2.5D GPU, display controller for displays up to 1024x768 and multiple microphone inputs for far-field audio capture. An acoustic activity detection function remains operational while in deep sleep to support low-power wake word detection.

The MCUs offer large memories, extensive analog functions, and various communication interfaces including I3C, USB, and optional CAN-FD and ethernet.

PSOC Edge MCUs can be used with Infineon’s DEEPCRAFT™ AI Suite. Central to the suite is DEEPCRAFT Studio, an end-to-end AI and ML development platform that integrates with ModusToolbox to reduce time to market and the complexity of edge AI development, regardless of ML expertise. DEEPCRAFT Studio supports data collection, analysis, and processing to build and train models. Users can also import their own proprietary models or use open source models. Regardless of the model source, developers can use DEEPCRAFT Studio to validate and optimize the model for use on PSOC Edge Devices, and then finish development and deployment in ModusToolbox.

Two development kits are available for PSOC Edge MCUs. The PSOC Edge E84 AI Kit is a low-cost kit with multiple onboard sensors for fast prototyping and evaluation of the MCU’s AI capabilities. The kit includes a compact board with the PSOC Edge E84 MCU, external memories, and an AIROC Wi-Fi plus Bluetooth combo chip. It also has a 60 gigahertz radar, microphones, and a barometric pressure sensor, plus a separate camera module. It comes pre-programmed with a voice assistant demo and a gesture detection demo.

The PSOC Edge E84 Evaluation Kit is a general purpose evaluation kit with full functional integration of all interfaces. The kit consists of a baseboard and SOM. The SOM includes the PSOC Edge E84 MCU, external memories, a PSOC 4000T as a CAPSENSE coprocessor, and an AIROC Wi-Fi and Bluetooth combo chip. The baseboard includes microphones, sensors and connectors for Ethernet, USB and other interfaces. The kit also includes a 4.3 inch capacitive touch display, a camera module, and multiple ML code examples.

Infineon PSOC Edge MCUs are intended to enable more innovative and user-centric designs in edge devices across many industries including consumer and industrial robotics, human machine interfaces in white goods and factory automation, wearables, smart home devices, and security cameras.