In May 2025, Unitree's humanoid robot appeared at the world's first combat competition exclusively for humanoid robots—the CMG World Robot Fighting Championship—demonstrating precise and professional fighting combinations.
For the competition, Unitree used CHINGMU's optical motion capture system to provide key technical support for robot action training and performance optimization. The system's high precision and low latency capture full-body joint trajectories, enabling subtle posture tuning and in-depth analysis of complex actions.
During pre-competition training, CHINGMU's system provided high-quality training data collection for Unitree's humanoid robot, simulating various complex actions in competition scenarios and recording multi-modal motion data, helping the R&D team accurately evaluate motion precision, trajectory deviation, and posture stability, and optimize action algorithms accordingly. Supported by this system, Unitree's humanoid robot perfectly executed complex fighting and cooperative actions at the competition, demonstrating outstanding motion performance and helping the team achieve excellent results.
