How Infineon PSOC™ Edge Accelerates Edge AI Deployment
As generative AI continues to reshape how people interact with technology, market attention is shifting from cloud AI toward Edge AI. Across smart homes, wearable devices, industrial monitoring, and robotics, more applications are bringing AI capabilities directly to end devices to achieve faster response times, lower power consumption, and stronger data privacy.
Deploying AI models on a microcontroller (MCU), however, is far from simple. Developers must balance computing performance and power consumption while also addressing memory capacity, sensor integration, security, and software development complexity.
To meet these needs, Avnet and Infineon have introduced the PSOC™ Edge E84 Edge AI Solution. Combining a complete hardware platform with AI development tools and a software ecosystem, the solution helps developers rapidly build the next generation of intelligent end products.
A Next-Generation MCU Purpose-Built for Edge AI
At the At the core of the solution is Infineon’s next-generation PSOC™ Edge E84 MCU, which integrates high-performance computing, a low-power architecture, and dedicated AI hardware acceleration. PSOC™ Edge E84 combines an Arm® Cortex®-M55 processor with Helium™ DSP extensions, an Arm® Ethos™-U55 NPU, and the Infineon NNLite AI accelerator, enabling more complex AI inference workloads on an MCU platform.
Compared with traditional MCUs, PSOC™ Edge E84 not only delivers greater AI processing capability but also expands the range of Edge AI applications that can run on end devices. Its dual-core Cortex-M55 and Cortex-M33 architecture addresses both application processing and system management requirements. Together with the Ethos™-U55 NPU and NNLite AI accelerator, it maintains strong energy efficiency during AI inference. Built-in security and privacy protection capabilities further support the development of next-generation intelligent edge devices with contextual awareness.
Multimodal Sensing for Intelligent Applications
The value of Edge AI begins with sensing.
Avnet’s PSOC™ Edge E84 Edge AI Solution integrates multiple sensing components and communication interfaces, including:
- 60 GHz radar sensor
- 6-axis IMU (inertial measurement unit)
- 3-axis electronic compass
- Temperature and humidity sensor
- Analog and digital MEMS microphones
- Wi-Fi 6/6E and Bluetooth® wireless connectivity
- External PSRAM and Flash memory expansion
By coordinating these sensors, the system can process audio, motion, position, environmental information, and visual data, providing a foundation for sensor fusion and multimodal AI applications.
In a smart home, for example, a device can use radar to detect human presence or combine local voice recognition with on-device voice interaction. In industrial environments, vibration, audio, and environmental sensor data can be fused to support equipment health monitoring, anomaly detection, and predictive maintenance.
Supporting a Complete Range of AI Applications, from Voice to Vision
As Edge AI algorithms continue to advance, end devices are moving beyond basic sensing functions.
The PSOC™ Edge E8x family supports a wide range of AI application scenarios, including:
- Presence detection
- Gesture recognition
- Motion detection
- Predictive maintenance
- Anomaly detection
- Autonomous operation
- Voice AI interaction
- Vision-based AI applications
Voice AI is one of the most important areas of Edge AI development today. PSOC™ Edge AI supports wake-word detection and can also perform voice activity detection (VAD), Voice ID, and local voice interaction, delivering real-time and secure voice experiences on the device.
A One-Stop Workflow from Model Training to Deployment
For many developers, deploying an AI model can be more challenging than building the model itself. Infineon addresses this challenge with a complete Edge AI ecosystem:
ModusToolbox™
Provides a complete development environment, device configuration tools, middleware, and extensive code examples to accelerate embedded product development.
DEEPCRAFT™ AI Suite
Supports the complete workflow for data collection, preprocessing, model development, training, deployment, and optimization.
DEEPCRAFT™ Ready Models
Offers ready-to-use AI models that developers can apply directly, including:
- Gesture recognition
- Snore detection
- Sound direction detection
- Cough detection
- Factory alarm detection
- Baby cry detection
- Fall detection
- Siren detection
Through the DEEPCRAFT ecosystem, developers can accelerate AI model development, deployment, and validation, helping shorten product development cycles and time to market.
An Edge AI Architecture Balancing Performance, Security, and Low Power
As AI moves into end devices, security becomes increasingly important.
The PSOC™ Edge E8x family supports Infineon Edge Protect Category 4 (EPC4) security architecture and PSA Certified Level 4 capabilities, providing features such as secure boot, key protection, and a secure execution environment.
Combined with the low-power Cortex-M33 subsystem and NNLite AI accelerator, the system can continuously execute critical sensing and AI tasks in ultra-low-power modes.
This makes the PSOC™ Edge E8x family of MCUs particularly well suited for intelligent sensors, battery-powered devices, and industrial edge gateways that require extended operating time.
Conclusion: Closing the Gap from Edge AI Concept to Product
As generative AI and edge intelligence continue to evolve, the market needs more than increased computing performance. It needs a complete platform that balances performance, power consumption, security, and development efficiency.
Avnet’s Infineon PSOC™ Edge E84 Edge AI Solution is designed to meet this need. By combining PSOC™ Edge E8x family MCUs, a multimodal sensing platform, the DEEPCRAFT AI ecosystem, and comprehensive development support, the solution helps developers rapidly create a new generation of Edge AI applications for smart homes, industrial monitoring, robotics, wearable devices, and intelligent human-machine interfaces.
When AI can sense, understand, and make decisions in real time on end devices, truly intelligent experiences will extend from the cloud to every edge node.
Newsletter
Discover what’s trending in the channel and latest products and solutions with Avnet Asia newsletter.
Learn MoreContact Avnet Design Services
Stay Tuned With Us. Stay up to date with our latest news on Reference Solutions.
Contact Us