Bringing RoboCup-honed technology and a spirit of challenge to the real world.
With a core of members active at the RoboCup world championship in service robotics, AIRS works on research, development, and real-world deployment in robotics and AI.
About AIRS
From curiosity to implementation.
We learn about new research and technology early and try it with our own hands. From AI to humanoids, we expand what we can do through repeated implementation and validation.
Services
Robotics and automation
We automate transport, assembly, and inspection with robots, from development and deployment to operation and maintenance.
AI and software development
We develop and maintain AI and machine learning software tailored to your operations.
Research and technical support
In robotics and AI, we provide contract research and technical consulting, and support the development of engineers and students through education and training.
Research & Development
Achieving point-cloud mapping and navigation over 100 m × 100 m with Unitree Go2, without additional hardware
The LiDAR built into Unitree Go2 could not produce an accurate map on its own. Our own system solved this, achieving point-cloud mapping and navigation over a 100 m × 100 m area without additional hardware.
Teaching a dual-arm robot to fold a handkerchief
Trained on 100 demonstrations, ACT folded a handkerchief into quarters in 37 of 40 trials under the same cloth and placement conditions. Failures with different patterns or positions show the limits we are investigating on the real robot.
From Replicating RL Token to Reward Design with Humans and AI
An experiment teaching a robot to move oranges onto a plate revealed that its success metric measured proximity instead. Human video review led to revised success criteria, rewards, and AI audit questions. Here is what was verified and what remains unresolved.
Grasping, carrying, and placing for RoboCup with an LLM and rule-based control
Using household tasks from RoboCup, we grasp, carry, and place objects with rule-based control, and use an LLM to update those rules based on test results in simulation.
News
Our representative, Yuta Oriike, presented a poster at the iHR workshop
At the 15th Intelligent Home Robotics workshop (iHR15) on September 19, 2026, we presented a reproduction study of RL Token and our work on reward design by people and AI.
We launched our official X and YouTube accounts
We will share our research and development activities and robots in action.
Social
We share our daily activities and robots in action on X and YouTube.
X
@airobot_systemsYouTube
@airobot_systems