Daily Technology
·23/03/2026
A significant milestone in humanoid robotics has been achieved as China's Unitree G1 robot demonstrated the ability to sustain a tennis rally with a human player. The robot, developed by Galbot Robotics, showcases advanced capabilities in real-time reaction and whole-body coordination, marking a notable progression in the field.
The G1's performance is powered by Galbot's proprietary Latent system, which the company describes as the world's first real-time whole-body planning and control algorithm. This system enables the robot to execute millisecond-level reactions, strike the ball with precision, and move with natural, fluid motions across the court. The complexity of tennis, with its high-speed ball travel and unpredictable trajectories, serves as a demanding benchmark for testing a robot's perception, control, and decision-making faculties under dynamic conditions.
This development represents a fundamental shift from earlier robotic capabilities, which often focused on pre-programmed mechanical motion replication. In contrast, the G1's ability to engage in a rally demonstrates an intelligent, decision-driven response to a constantly changing environment. In simulation tests, the Latent system achieved a forehand shot success rate of up to 96%. When deployed on the physical Unitree G1, it successfully maintained rallies by consistently returning the ball to the opponent's side of the court.
According to industry analysts, this breakthrough suggests a broader potential for humanoid robots beyond specialized tasks. As these machines become more adept at handling complex, scenario-based interactions, their application could expand into households and key sectors like industry, agriculture, and manufacturing. The ability to learn and adapt in real-world situations could lower training costs, improve scalability, and accelerate the adoption of robotics across various service and automation industries. While the current focus is on tennis, the underlying system is adaptable for other tasks where high-quality human motion data is limited, signaling a new era for intelligent robotics.









