In GALBOT's research on humanoid robot high-dynamic tennis rallying, CHINGMU's high-precision motion capture system provides core data support for robot skill training and algorithm optimization. The system features high dynamic response, precisely capturing human tennis players' high-speed actions such as swings, runs, and hits, while recording multidimensional data including limb joint motion, racket trajectory, and impact force, helping construct robot motion skill models.
The diverse human tennis action data captured by CHINGMU's system provided key support for the LATENT research framework jointly proposed by GALBOT and Tsinghua University, enabling robots to break free from pre-programmed limitations, autonomously judge incoming balls, adjust footwork, and complete continuous returns. The system's low latency and high precision ensure real-time and accurate data acquisition, helping robots quickly learn natural and fluid striking motions through reinforcement learning, achieving a breakthrough in humanoid robot high-dynamic tennis rallying.
