Graduate School of Engineering, Nagoya UniversityComputer Architecture Group
From Silicon to Robotics.
Redefining computer architecture through embodied intelligence
About
Hardware for AI, and AI for hardware —
all the way to robots that move in the real world.
As AI evolves at an explosive pace, conventional computing systems are reaching a major turning point. We work from two directions — chips that run AI fast and efficiently, and AI that automates complex chip design — and vertically integrate the results into real robotic systems. From algorithms to FPGA and ASIC implementation and real-robot control, our research bridges software and hardware end to end.
Research
View all
AI for Circuits
Automating IC design with LLMs
An EDA flow in which LLM agents go from natural-language specifications to Verilog and layout (GDS) on their own. The LLM repeatedly revises the layout while reading DRC/LVS results.
Circuits for AI
High-efficiency edge AI accelerators
SRAM compute-in-memory fabricated and measured in 22 nm, multiplication-free LUT-based accelerators, spiking neural networks — domain-specific architectures that run AI with maximum efficiency.
Systems & Robotics
Robotic applications of our own chips
Low-latency bilateral control on FPGA, faster motion generation for VLA models, and imitation learning with haptics. We deploy the chips and circuits we design onto real robots.
Contact
For collaboration or technical consultation, see "For Companies".
For lab visits or graduate admission, please contact the faculty directly.