Back to Rainbow Robotics
RB-Y1

RB-Y1

RB-Y1

Last updated: October 11, 2026
Locomotion Type
Wheeled
Stage
Mass Production
Height
140 cm
Degrees of Freedom
20 DOF
Weight
131 kg
Max Speed
5.4 km/h
Dual-Arm Payload
6 kg
Release Year
2024
Computing
A dual-PC architecture that deliberately separates real-time control from user applications. The Robot PC handles physical motion and safety, while the User PC runs ROS 2, the RB-Y1 SDK and user applications, so high-performance control and flexible algorithm development coexist rather than competing. The control structure is left open on purpose: the robot does not prescribe a fixed control method, and the SDK lets researchers design and verify optimal control, impedance control and whole-body control for teleoperation, human-robot interaction, real-world manipulation and AI learning. A research pipeline ties simulation, teaching and the robot together, with a URDF, MJCF and USD-based model for pre-validating control algorithms and operational strategies, then Learning from Demonstration data collected through a leader arm or VR-based teaching and applied to the robot through the same SDK interface. The system runs on 48 VDC, its exterior is aluminium, and the drive train inherits actuator technology verified on the company's collaborative robots along with a joint brake structure that holds posture when power is cut off.

The RB-Y1 is Rainbow Robotics' humanoid research platform, and the company is careful about what that means: it says outright that this is not a finished robot for one task but a mobile dual-arm platform for experimenting with manipulation, mobility and perception in real physical environments, built for AI manipulation learning, teleoperation and whole-body control research. The design decisions follow from that. It stands 1.4 m tall on a 600 by 690 mm footprint and weighs 131 kg in its standard configuration, with 20 degrees of freedom across two wheels, six-degree-of-freedom legs, two seven-degree-of-freedom arms and a two-degree-of-freedom head; a four-wheel configuration raises that to 26 DOF and 170 kg. Each arm lifts 3 kg and reaches 600 mm to the wrist before the hand, with plus or minus 0.05 mm repeatability inherited from the company's collaborative robots. The wheels are the point: it treats mobility and manipulation as one integrated task rather than two, and drives at 1.5 m/s. Underneath sits a dual-PC architecture that keeps real-time control and safety on the Robot PC while the User PC runs ROS 2, the RB-Y1 SDK and user code, and the control structure is left deliberately open so researchers can build optimal, impedance or whole-body controllers rather than accept a fixed one. The research loop is closed properly: a URDF, MJCF and USD model lets algorithms be validated in simulation first, then Learning from Demonstration data is collected through a leader arm or VR teaching and pushed to the robot through the same SDK interface. Power is 48 VDC with a 1,250 Wh or 2,500 Wh battery, the exterior is aluminium, and a joint brake structure holds posture even when power is cut, which matters for long repeated experiments.

Application Scenarios

Robotics research and expansion, specifically AI-based manipulation learning, teleoperation, human-robot interaction, whole-body control research and impedance control validation; research institutes and specialist markets needing a development platform; and real physical environments where manipulation, mobility and perception must be tested together rather than in isolation.

Subscribe to Weekly Report

Get weekly updates on humanoid robot funding, new companies, and product launches.