Robot Soccer Misjudgments at Digital Trade Expo Challenge Show Imperfection as Part of Competition

According to Leaderobot, on September 25 at the 5th Global Digital Trade Expo, the 2026 Global Digital Trade Innovation Competition's Embodied AI Robot Challenge was in full swing. Losing balance during a high-speed run, making a misjudgment in robot soccer, or stopping to replan a route when facing an obstacle—these seemingly imperfect moments became part of the competition itself.

The event ran from September 23 to 27 at the Hangzhou Grand Convention and Exhibition Center as a side event of the 5th Global Digital Trade Expo. It attracted more than 100 teams from domestic universities and technology companies, with 14 official events covering speed, competition, and endurance. Organizers did not treat mistakes as flaws to be hidden; instead, they used them as a window into robots' real capabilities.

In terms of event design, the 60-meter and 100-meter sprints, shuttle runs, and field obstacle races mainly test robot motion control, balance, and the ability to navigate complex environments. Set-shot basketball, 3v3 soccer, free fighting, and tug-of-war further test perception, decision-making, and whole-machine coordination in dynamic environments.

The sprint events are not merely about straight-line speed. A robot that runs quickly in a clear lane may still struggle when it must turn, avoid an obstacle, or recover from a slip. The obstacle race adds uneven surfaces and unexpected barriers, demanding rapid replanning and stable locomotion under pressure.

Set-shot basketball and 3v3 soccer introduce moving targets and opponents, where perception and decision-making must operate under time pressure. Free fighting and tug-of-war push joint torque and structural integrity to their limits. Tug-of-war, in particular, tests sustained force and balance against an external load.

From running and kicking a ball to competitive combat, what is being compared is not just how many actions a robot can perform, but whether those actions can be completed stably, accurately, and continuously. Misjudgments in soccer confrontations and pauses to replan routes when facing obstacles might be edited out in the lab, but they are presented in full on the competition floor.

Compared with isolated tests in a laboratory, a high-intensity, continuously operating, and variable-filled competition environment is closer to the real work scenarios robots may face in the future. The event also included product displays and industry exchange sessions, aiming to bring test data accumulated on the field and engineering problems exposed during competition back to the R&D side. The goal is to push embodied AI from being able to move toward being able to work—and work reliably—in real scenarios.

Fighting events test the body's peak joint torque, balance recovery under impact, structural strength, and stability during continuous operation. Speed and endurance races test motion control, battery management, and heat dissipation. Soccer confrontations test collaborative decision-making among multiple robots. These are not abstract metrics; they determine whether a robot can keep working after an impact, sustain performance over long periods, or coordinate with others without human intervention.

When these capabilities are transferred to industrial and service scenarios, they correspond to how long a robot can work continuously and stably in a real environment. In a factory, for example, a robot may need to maintain precision after repeated shocks, manage its own power and heat, and avoid collisions with other machines. In a service setting, it may need to perceive unpredictable people and objects and adjust its plans in real time.

Treating imperfection as part of the competition corresponds to the industry's sober judgment of the current technological stage. Robots are still far from being able to stably complete all tasks; the point of the competition is to expose problems, not to display perfection. In continuous combat and high-speed motion, any misjudgment, loss of balance, or route replanning becomes direct material for testing the reliability of the whole system.