Unitree G1 EDU | U1-U14 | 人形机器人开发平台

G1 EDU Standard | 23 DOF | basis of the Professional variants --> Dummy Hand | EDU-U1
G1 EDU Plus | 29 DOF | basis for Ultimate variants --> Dummy Hand | EDU-U2

G1 EDU- Professional (23 DOF Basis)

• Professional A → Dex3-1 | 3 Finger | EDU-U9

• Professional B → Dex3-1 | 3 Finger with tactile sensors | EDU-U10

• Professional E → BrainCo Revo 2 Basic | 5 Finger | EDU-11

• Professional F → BrainCo Revo 2 Touch | 5 Finger with tactile sensors | EDU-12

• Professional G → Dex2/5 | 5 Finger | EDU-14

• Professional H → Dex1-1 Gripper Plus

• Professional I → Wuji Hand 2 | 5 Finger

G1 EDU- Ultimate (29 DOF Basis)

• Ultimate A → Dex3-1 | 3 Finger | EDU-U3

• Ultimate B → Dex3-1 | 3 Finger with tactile sensors | EDU-U4

• Ultimate C → Inspire DFQ | 5 Finger | EDU-U5

• Ultimate D → Inspire E2 | 5 Finger | with tactile sensors | EDU-U6

• Ultimate E → BrainCo Revo 2 Basic | 5 Finger | EDU-U7

• Ultimate F → BrainCo Revo 2 Touch | 5 Finger with tactile sensors | EDU-U8

• Ultimate G → Dex2/5 | 5 Finger | EDU-13

• Ultimate H → Dex1-1 Gripper Plus

• Ultimate I → Wuji Hand 2 | 5 Finger

The Unitree G1 EDU U3 is a 1,270 mm tall humanoid development platform for research, education, and application-oriented robotics projects. Weighing around 35 kg, the system remains manageable for laboratory setups. The replaceable 13S smart battery provides approximately two hours of operation, depending on motion and computing load. A depth camera and 3D LiDAR provide environmental data for perception, navigation, and autonomous experiments. Dual encoders at the joints, internal cable routing, and local air cooling support repeatable motion. Knee joints with up to 120 Nm of torque and an arm payload of up to 3 kg provide additional capacity for dynamic motion sequences and mobile manipulation. Performance limits must be taken into account when designing the setup, motion profiles, and payload.

The G1 EDU U3 comes with force-controlled Dex3-1 three-finger hands for grasping and manipulation experiments. Each hand has seven degrees of freedom: three in the thumb and two each in the index and middle fingers. This enables the development of different grasping patterns and object-specific motion sequences. Each hand weighs approximately 650 g, measures 175 × 88 × 77 mm, and connects via USB. The hand operates without tactile sensors; forces must therefore be controlled through the control system, known object characteristics, and appropriate motion limits. Wrist-mounted RGB cameras with 1,920 × 1,080 pixels and a 130° diagonal field of view provide additional image data for object detection, grasp point detection, and vision-guided manipulation.

The G1 EDU platform provides high-level and low-level interfaces for custom applications. Developers can use these interfaces to implement motion sequences, autonomous functions, and joint-level control tasks. The integrated expansion dock with up to 100 TOPS of computing power is suitable for AI algorithms, machine vision, navigation, and positioning directly on the robot. Wi-Fi 6 and Bluetooth 5.2 facilitate integration into laboratory environments, while Smart OTA supports system updates. Development documentation and ecosystem support make it easier to get started. Custom grasping sequences for the Dex3-1 hands can be linked with camera data. Control cycles, force limits, collision zones, and safety functions must be configured for the specific project and tested before operation.

The G1 EDU U3 is suitable for secondary development projects that combine perception, motion, and grasping. Typical areas include humanoid locomotion, balance control, AI-based object detection, teleoperation, and mobile manipulation. The Dex3-1 hands enable three-finger grasping experiments, while the wrist cameras provide image data close to the working area. Compared with models without actuated hands, the U3 is more focused on vision-guided manipulation. Variants with tactile Dex3-1 hands additionally capture contact information through 33 sensors per hand. Other G1 EDU models use five-finger hands, tendon-driven hands, or parallel grippers. Before ordering, the number of degrees of freedom, hand version, cameras, and scope of delivery should be confirmed.

RBTX supports users with commissioning and integrating the Unitree G1 EDU U3 into a functional experimental or demonstration setup. This includes training for operation, programming, and safe testing procedures, as well as selecting additional sensors, computing hardware, and safety components. For applications using the Dex3-1 hands, RBTX can coordinate grasping objects, motion ranges, camera positions, and software interfaces. During the integration project, the robot, hand control, machine vision, network, and application are considered as a complete system. This allows the required functionality, responsibilities, and acceptance criteria to be defined at an early stage. Force limits, collision zones, the emergency stop concept, and safe retreat motions are taken into account during the concept phase.