LUS2
LUS2 is a humanoid research robot by Lumos Robotics designed for safe interaction, manipulation testing, and applied development.
Description
Key Features
Ultra-Fast Recovery and Agility
Springs from lying flat to standing in 1 second using neural motion planning and high-torque servos (360 N·m max joint torque), enabling dynamic balance in unstructured environments.
Advanced Multimodal Sensing
Intel RealSense D435i, 360° LiDAR, IMU, tactile sensors (finger/palm, 0.05mm accuracy), supporting SLAM navigation, precise grasping, and environmental perception.
High-Performance Compute
NVIDIA Orin AGX (275 TOPS) with dedicated real-time controller, Linux OS, ROS2, LLM APIs for AI-driven autonomy and custom development.
Dexterous Manipulation
28 DOF total, 10 DOF hands with 5 fingers, 5kg payload, visual-tactile grippers (LUS GT) for complex object handling.
Modular Open Architecture
Open SDKs/APIs, EtherCAT/Wi-Fi connectivity, easy integration for research, education, and industry applications.
Safety and Reliability
Force-limiting joints, collision detection, emergency stops; robust for human collaboration, 3-hour runtime, -10°C to 55°C operation.
Specifications
| Availability | Early production / limited release |
|---|---|
| Nationality | China |
| Website | https://www.lumosbot.tech/ |
| Degrees Of Freedom, Overall | 38 |
| Degrees Of Freedom, Hands | 10 |
| Height [Cm] | 160 |
| Manipulation Performance | 3 |
| Navigation Performance | 3 |
| Max Speed (Km/H) | 4 |
| Strength [Kg] | 5 kg whole‑hand capacity |
| Weight [Kg] | 65 |
| Runtime Pr Charge (Hours) | 3 |
| Safe With Humans | Yes |
| Cpu/Gpu | Industrial CPU + optional GPU |
| Camera Resolution | RGB + depth, ~1080p |
| Connectivity | Ethernet, USB, Wi‑Fi |
| Operating System | Linux-based ROS environment |
| Llm Integration | Possible via external system integration |
| Latency Glass To Action | 150–250 ms |
| Motor Tech | Electric servo motors |
| Gear Tech | Harmonic / planetary gears (assumed) |
| Main Structural Material | Aluminum alloy + composites |
| Number Of Fingers | 5 per hand |
| Main Market | applied humanoid R&D, Research & education |
| Verified | Not verified |
| Walking Speed [Km/H] | 3.5 |
| Shipping Size | ~180 × 80 × 65 cm |
| Color | Black |
| Manufacturer | Lumos Robotics |
| Dimensions | 160 cm height, ~50x30 cm width/depth (shipping: ~180x80x65 cm) |
| Weight | 57 kg |
| Dof | 28 total (6 DOF/leg, 10 DOF/hands) |
| Processor | NVIDIA Jetson Orin AGX (275 TOPS), dedicated real-time control processor |
| Os Software | Linux-based OS, ROS2 compatible, open APIs/SDK (Python/C++) |
| Sensors | vision: Intel RealSense D435i (RGB + depth, ~1080p), proprioception: IMU, gyroscope, force/temperature sensors, ultrasonic, tactile: Finger/palm sensors (0.05mm accuracy, 30fps, CAN/UART/EtherCAT), navigation: 360° LiDAR, indoor/visual SLAM |
| Actuators | Self-developed electric servo motors (e.g., Lumos-60-30: 550g, 64x58.5mm, 22.8Nm rated/102Nm peak torque, 140 RPM @48V, dual absolute encoders) |
| Gears | Planetary reducer (reduction ratio 30) |
| Materials | Aluminum alloy + composites |
| Battery | 10 Ah (10000mAh), 54V DC, 3h runtime, 2-3h charge |
| Power Thermal | 24-70V operation (48V recommended), -10°C to 55°C environment |
| Connectivity | Wi-Fi, Ethernet, USB, EtherCAT/CAN/UART |
| Performance | Max speed 2 m/s, joint torque 360 N·m, latency 150-250 ms, payload 5 kg |
| Safety | IP N/A, force-limiting, collision detection, e-stop |
Curated Videos