XGO-mini2SW is a reinforcement-learning wheeled-legged robot dog with a Raspberry Pi CM5 brain, VLM perception, LiDAR, and a 4-DoF arm.
- Reinforcement Learning Locomotion: Adapts to changing terrain, maintains balance, and recovers from disturbances in real time with an open-source Sim-to-Real deployment pipeline.
- Wheeled-Legged Mobility: Wheels for speed on smooth surfaces and 12-DoF legs for slopes and obstacles; FOC hub-wheel module switches configurations in minutes.
- Raspberry Pi CM5 + VLM + LiDAR: Recognizes objects, understands scenes, describes what it sees, and tracks targets with real-time spatial perception.
- 4-DoF Robotic Arm & Natural-Language AI: Interacts with objects; AI understands the codebase, writes and debugs software, and deploys directly from intent to motion.
- Fully Open-Source (Blockly, Python, C++, ROS): Built by LuwuDynamics from HIT robotics research, with a personalized AI companion layer and long-term memory.
