Enjoy AI &

Build your own AI !

FPGA2I Shield for Arduino


Notifications

SEMINARS

  • Notice
    Seminar: FPGA2I Shield for Arduino 2021 will be held after next spring (influence of new Corona Virus COVID19) [announced 2020/09/28]
  • Notice
    Grand Challenge 2020 (online) @ Hokkaido University [held 2020/June]
  • Notice
    Seminar 2: FPGA2I shield for Arduino postponed (influence of new Corona Virus COVID19) [announced2020/02/24]
  • Link
    Osaka University Seeds2019 First Step Training [held 2019/08/26,27]

See more

EXHIBITIONS

See more

PUBLICATIONS

See more

PRODUCTS

  • Notice
    Product launch: Details coming soon [First half of 2021]
    AI Circuit Lab: AI Shield for real-time learning and inference
  • Notice
    FPGA2I Shield Version3 Version with camera interface [released 2019/03/17]
  • Notice
    FPGA2I Shield Version2: Improved analog input/output unit [released 2018/08/04]
  • Notice
    FPGA2I Shield Version1: [released 2018/03/23]

Notice

  • Notice
    Product launch: Details coming soon [First half of 2021]
    AI Circuit Lab: AI Shield for real-time learning and inference
  • Notice
    Published in the Japanese technical magazine "Transistor Technology", December 2020 issue:
    Short term serialization: AI line tracer using RNN suitable for time series processing [published 2020/December]
  • Notice
    Published in the Japanese technical magazine “Transistor Technology”, November 2020 issue: New product announcement (AI Circuit Lab):
    Short term serialization: Real-time processing of learning and inference ! Introduction to Hardware AI [published 2020/November]
  • Link
    Maker Fair Tokyo 2020 Twitter [held 2020/10/4]

See more


About FPGA2I Shield

AI accelerator: Anyone can easily experience AI learning and
inference on a single board! !

View Movie

Pinch out to enlarge.
About User-driven AI Platform
A platform that can share information about AI hardware and software created on our Web Support Forum.
Join us, experiment with AI, and create your own AI.

Learning Example

Learn “Line”
(Line Predictor)
AI learns a shape of "line". It memorizes the shape using Recurrent Neural Network (RNN) suitable for sequence pattern learning.

More details

Learn “Difference” (XOR)
AI learns the “differences” between two items with exclusive OR (XOR).

More details

Learn “3 Colors”
AI learns the differences between the three "colors" by training with a neural network with three inputs.

More details


Specifications

(1)FPGA2I Shield V3 (Engineering sample)
The calculation of AI is heavy, and the Arduino that houses a small microcomputer does not have sufficient power. The FPGA2I Shield works as an AI accelerator and supports learning and inference online.

NameFPGA2I Shield Version3
FunctionAI Accelerator
FPGAIntel Max10: 10M25SCE144, ROM
Memory1Mbit SRAM 2pcs
InterfaceArduino pin compatible(5V)
μSD Card(for Configuration)
Camera I/F 20pins
Shield Size110mm × 53mm
System ControlArduino IDE(Sketch)
AI Design ToolsAI Circuits:Verilog
AI Config: Quartus Lite IDE⇒SD Card

Download Specifications

  • Caterpillar robot (Zumo Robot)
  • Operation with box
  • Connection b/w FPGA2I Shield & Arduino

(2)AI Circuit Lab (on sale in first half of 2021)
It is a compact AI shield with a size of about 19 mm x about 60 mm. It will be easier for you to incorporate it into the commodities, robots, tools, etc. that you usually use, and you can enjoy installing AI in a more familiar place.

It is scheduled to go on sale in the first half of 2021. (The photo uses a prototype)

NameAI Circuit Lab
FunctionAI Accelerator
FPGAIntel Cyclone10LP:10CL016YM164
Memory1Mbit SPI SRAM 2pcs
InterfaceArduino Nano33 pin compatible
μUSB (for configuration)
Camera I/F 20pin
Shield Size(60mm × 19mm)
System ControlArduino IDE(Sketch)
AI Design ToolsAI Circuits:Verilog
AI Config:Quartus Lite IDE →USB Cable

DL Spec(under construction)

  • Caterpillar robot (Zumo Robot)
  • Operation with box
  • Connection(+Arduino Nano33)

Partner companies/organizations

  • e-trees.Japan
  • トランジスタ技術
New member registration
New member registration is currently under preparation.
For more information, please email “support@fpga2i.org (formerly: fpga2i-support@ist.hokudai.ac.jp)."