Unitree – Go2 EDU Plus QUGV

Unitree – Go2

 

 

Unitree – Go2

Our center’s Unitree Go2 EDU Plus is a lightweight, agile quadruped robot integrated into the TRACE testbed as a high-mobility legged platform for edge AI and autonomous systems research. Weighing approximately 15 kg with battery, the Go2 EDU Plus (model Go2 Edu-U4) is equipped with a 100 TOPS NVIDIA Jetson Orin NX compute module mounted in its docking station, enabling onboard SLAM, real-time inference, and sensor fusion at the edge. Its sensor payload includes a Hesai XT16 3D LiDAR for high-fidelity point cloud mapping and navigation, a forward-facing Intel RealSense D435i depth camera for close-range depth perception, and a wide-angle HD camera for visual monitoring. Connectivity via WiFi 6, Bluetooth 5.2, and built-in 4G/eSIM supports both lab-network integration and field deployments. With a long-life 15,000 mAh battery providing 2–4 hours of runtime and a 33.6V/9A fast charger, the Go2 EDU Plus is suited for extended research sessions. Its open SDK supports ROS/ROS 2, Python, and C++ for both high-level application control and low-level torque and gait control, making it a versatile platform for locomotion research, multi-robot coordination, and perception experiments within the TRACE testbed.

Key Capabilities

  • Legged quadruped mobility; operates on uneven terrain, stairs, slopes up to 40°, and cluttered indoor/outdoor environments up to ~16 cm step height
  • Onboard perception: Hesai XT16 3D LiDAR, RealSense D435i depth camera, and wide-angle HD camera for SLAM, obstacle avoidance, and object detection
  • High-performance edge compute: 100 TOPS NVIDIA Jetson Orin NX for real-time inference, mapping, and autonomy stack execution
  • Multi-modal interaction: built-in microphone, speaker, 3W lighting, and voice recognition module for HRI experiments
  • Wireless connectivity: WiFi 6 (802.11ax), Bluetooth 5.2/4.2/2.1, 4G/eSIM for field and lab deployments
  • Full secondary development support: Python/C++ SDK with low-level torque control, foot-end force sensors, and ROS/ROS 2 compatibility
  • ISS 2.0 intelligent side-follow and wireless vector positioning for multi-agent coordination scenarios.

Documentation & Helpful Links

Hardware Configuration

Base Platform

  • Unitree Go2 EDU Plus (Go2 Edu-U4); gross weight ~15 kg (with battery); dimensions 70 cm × 31 cm × 40 cm
  • 12 degrees of freedom; payload capacity ~8 kg (limit ~10 kg)
  • Maximum speed: 0–3.7 m/s; maximum climbing slope: 40°; maximum step height: ~16 cm
  • Joint range: body ±48°, thigh −200° to +90°, calf −156° to −48°
  • Internal knee joint alignment with heat pipe assisted heat dissipation

Onboard Computer

  • Docking Station with 100 TOPS NVIDIA Jetson Orin NX — supports onboard AI inference, SLAM, perception pipelines, and ROS 2 node hosting
  • 8-core high-performance base CPU

Sensor Payload

  • 3D LiDAR: Hesai XT16 (high-end), with navigation algorithms and technical support
  • Depth Camera: Intel RealSense D435i (forward-facing)
  • RGB Camera: Wide-angle HD camera (forward-facing)
  • Foot sensors: onboard foot-end force sensors for terrain and balance research

Communications

  • WiFi 6 dual-frequency (802.11ax)
  • Bluetooth 5.2 / 4.2 / 2.1
  • Built-in 4G / eSIM with intelligent OTA upgrade support

Autonomy Architecture

  • ROS/ROS 2 compatible autonomy platform
  • Supports both high-level application-layer control (Python/C++) and low-level torque/gait control APIs
  • Full SDK access: raw motion, sensor, and telemetry data for custom autonomy development
  • APP-based graphical programming for rapid prototyping
  • Intelligent OTA upgrade for firmware and software updates
  • Compatible with Unitree’s open-source ecosystem on GitHub

Compliance & Procurement

  • Conforms to National Defense Authorization Act (NDAA) 2020 requirements for DoD UAS
  • Available for institutional purchase; coordinated through ARL and DoD procurement channels

Integration Notes for Multi-Robot Testbed

  • Networking: Assign static IP/hostname on lab VLAN; document DHCP reservations for VOXL 2
  • Time Sync: NTP/PTP time synchronization across all TRACE robots and logging PCs
  • Data Pipeline: Define ROS 2 topics to record (RGB, stereo, IMU, rangefinder, state estimates); set storage and retention policies
  • Interop: Bridge VOXL 2 ROS 2 nodes with ground robot stacks for air-ground teaming experiments
  • Safety Interlocks: Shared E-stop policy, geofenced flight areas, and speed/altitude caps in multi-robot zones
  • Calibration: Camera intrinsic/extrinsic calibration using shared calibration board; align coordinate frames with ground robot reference frames
  • Versioning: Track VOXL 2 OS, PX4 firmware, ROS 2 stack versions, and autonomy algorithm versions.