A collaborative team of researchers from Tsinghua University and Peking University has developed a groundbreaking system named LATENT, which allows the Unitree G1 humanoid robot to engage in high-speed tennis rallies. This innovative technology leverages "imperfect" human motion data to enhance the robot's performance, employing corrective residual actions to improve its agility and precision on the court. The project showcases advancements in robotics and artificial intelligence, aiming to bridge the gap between human-like movement and robotic capabilities. The research highlights the potential for robots to learn from human behavior, even when that behavior is not perfect, thereby expanding the applications of robotics in sports and beyond.
Leave a comment