AON1120 is a Ultra-Low-Power RISC-V SoC Designed for Always-On Edge AI Applications

AON1120 is a Ultra-Low-Power RISC-V SoC Designed for Always-On Edge AI Applications

1.198
Views
0 Comments

At CES2024, AONDevice unveiled its new on-device machine learning chip, the AON1120. It features a RISC-V core, two NPU coprocessors, and a DSP, and operates at a remarkably low power consumption of just 260μW for full-load audio classification tasks.

The AON1120 is a system-on-chip that includes a RISC-V core, a DSP, and two NPUs for machine learning. It features multiple modes, drawing only 80μW for audio classification in listening mode and less than 260μW during full AI processing.

In addition to voice commands, the AON1120 supports biometric, gesture, and motion recognition, along with sensor fusion. It has 16 GPIU pins and offers 40 adaptable output classifications. This versatility makes it ideal for health monitoring, sleep analysis, voice-activated controls, driver authentication, and industrial equipment monitoring.

The company is offering an evaluation board featuring the edge AI processor, a TDK T5838 microphone, a USB controller, and USB power, with measurement points for core and IO power. Each kit includes a “complete AI tool suite” for dataset augmentation, training, and optimizing inference.

Detailed Specification of AON1120 Ultra-Low-Power RISC-V SoC :

  • Chip Design:
    • Type: Sensor fusion chip.
    • Architecture: Edge AI, optimized for super low-power operations.
  • Processing Units:
    • Core: RISC-V core.
    • Neural Processing Units (NPU): Dual NPU for advanced AI tasks.
    • Digital Signal Processor (DSP): Specialized HW DSP for efficient data processing.
  • Energy Efficiency:
    • Full Processing Power: Less than 260μW for 100% AI task processing.
    • Listening Mode: Below 80μW, enhancing device battery life.
  • Interfaces:
    • GPIOs: 16x General-Purpose Input/Output pins for diverse application support.
  • Packaging:
    • Form Factor: Compact 40QFN, suitable for various device integrations.
  • Sensory Capabilities:
    • Voice Command Recognition: High reliability with contextual awareness.
    • Biometric Recognition: Advanced processing for secure identification.
    • Gesture and Motion Detection: Accurate interpretation for interactive applications.
  • Accuracy and Performance:
    • Background Noise Handling: Over 90% detection rate at less than 0dB SNR with a single microphone.
    • Output Classifications: Over 40 adaptable classifications for various AI tasks.
  • Application Suitability:
    • Health and Wellness Monitoring: For proactive healthcare.
    • Smart Home Automation: Intelligent triggers for security and efficiency.
    • Sleep Analysis: Tools for analyzing and improving sleep patterns.
    • Accessibility Aids: Hands-free control for accessibility.
    • Voice-Activated Remote Controls: Enhanced home entertainment experience.
    • Driver Authentication: Biometric speaker identification in vehicles.
    • Industrial Equipment Monitoring: Early fault detection with sound recognition.
  • Company Statement:
    • CEO Mouna Elkhatib’s Vision: Commitment to super low-power, cutting-edge edge AI for always-on applications.
  • Availability and Demonstrations:
    • CES 2024 Presence: Interactive exploration of the chip’s capabilities.
    • Collaboration Opportunities: Available for meetings and partnerships at CES 2024.
  • Further Information:

AONDevices will showcase their evaluation kit at CES 2024 in Las Vegas so at the time of writing pricing and availability details are yet to be announced.

Please follow and like us:
Pin Share
About Debashis Das

Tech Content Creator | Hardware Design Engineer | IoT Enthusiast | Digital Storyteller | Video Creator | From crafting digital dialogues at Semicon Media to orchestrating online engagement at Electronics-lab.com, my journey is about making tech talk interesting and accessible. In the quiet corners of the day, I advocate for easier tech education, one tutorial at a time. Let's connect and keep the digital dialogue going!

view all posts by das
Subscribe
Notify of
guest

0 Comments
Inline Feedbacks
View all comments
Get new posts by email:
Get new posts by email:

Join 97,426 other subscribers

Archives