A1 Unitree Robot

A1 robot dog from Unitree

Unitree A1 is a four-legged robot for research and development of autonomous systems in the areas of robot grid interaction (HRI), SLAM, and transportation. With four legs as well as 12 DOF, this robot can tackle a variety of terrains. The A1 is based on nearly 5 years of research and features an advanced drive and power management system that delivers speeds of 3.3 m/s or 11.88 km/h with run times of up to 2.5 hours. In addition, the motors have 33.5 N-m of torque, which also allows you to jump or somersault.

Multi-camera intelligent depth camera Multi-camera intelligent depth camera

Active infrared stereo depth technology.
The distance at depth ranges from 0.3 to 10 meters.
The resolution of the camera is 1080P.
Accuracy of error within 2 meters less than 2%

Maximum continuous running speed of 3.3 m/s (11.88 km/h).
The fastest and most stable small and medium-sized four-legged robot in China
(World record: 3.7 m/s achieved by MIT on a treadmill).
A speed of 3.3 m/s is close to adult running and can be used as a running partner.

Patented fuselage, with a layout structure and design

• Heavy-duty and lightweight housing design, simple and convenient maintenance.
• External input power interface 24V, external output power supply 5V, 12V, 19V, can be easily connected to external expansion devices.
  
• Standard high performance embedded dual master control (perception supervisor and motion supervisor).
• External interface: 4 USB ports, 2 HDMI, 2 Ethernet.
• Standard RealSense depth camera, building + obstacle avoidance.

Sensitive foot recognition

• The integrated force sensor on each foot is convenient for the robot to perceive the foot contact condition in real time in any state of motion.
• For better sensitivity and reliability, hardware force sensors and motor current measurement for foot force measurement are integrated into the program.
• The stem is waterproof and dustproof and easy to replace.

Vision-based autonomous following

• Track targets (people, certain moving objects) in visual range in real time.
• High real-time tracking, video update rate up to 25-30 fps.
• Support online learning of the following target characteristics with good reliability.
  
• The confidence level of the real-time tracking target is displayed by adjusting the color of the tracking field (green → red, 100% → 0%).