On March 15, Galaxy General introduced the world’s first real-time intelligent planning and control algorithm for humanoid robots, tailored for the complexities of tennis matches. This breakthrough equips the robot with the capability to engage in prolonged and dynamic tennis gameplay. Developed collaboratively by Galaxy General and a research team from Tsinghua University, the robot is designed to learn complex motor skills from imperfect human movement data, allowing it to execute tennis strokes and rallies with precision. Standing approximately 1.75 meters tall, the robot harnesses the LATENT algorithm to autonomously acquire tennis skills through deep reinforcement learning. Its advanced binocular vision system enables it to swiftly track incoming balls. In real-world tests, the robot achieved a remarkable 90.9% success rate in forehand strokes, consistently executing over 20 consecutive rallies while independently controlling ball placement and rhythm, with movements that closely mimic those of human athletes.