An autonomous badminton robot that can return 1,452 shots in a row
Shenchen Kaidong just built an autonomous robot practice partner called XA1 that predicts shuttle trajectories and repositions itself to return shots without any teleoperation. It handles clears, net drops, and smashes with an over 96% hit rate, even setting a Guinness World Record with 1,452 consecutive counter-hits. Why it matters: badminton is a brutal benchmark for embodied AI, proving robots can coordinate vision and fast mobility in real-time.
Robots assembling parts in factories or doing choreographed backflips on flat concrete are pretty common now. But put a robot on the opposite side of a badminton court, where a feather shuttlecock abruptly drops, spins, and changes speed, and most systems fall apart.
Shenchen Kaidong is tackling that exact nightmare with XA1, an autonomous badminton robot. The four-wheeled mobile manipulator operates completely without teleoperation, returning clears, drop shots, and even handling some smashes. Back in December 2025, it set a Guinness World Record by sustaining a mind-boggling 1,452 consecutive counter-hits, boasting an overall hit rate above 96%.
Splitting milliseconds to chase the shuttle
Badminton is not just about swinging a racket. The real struggle is coordinating vision, rapid footwork, and arm mechanics before the shuttle hits the floor.
To give XA1 lightning reflexes, the team ringed the court with over 30 spectral cameras filming at 1,000 frames per second. The onboard system recognizes the incoming shot and predicts its landing spot within 0.01 seconds, instantly translating trajectory data into wheel acceleration and racket articulation.
Hardware balance is equally tricky. The robot weighs roughly 50 kg; make it heavier for stability and inertia ruins agility, make it lighter and the racket recoil knocks it off balance. Even the tires required custom development to ensure high friction without wearing down from sudden stops.
Sparring with human champions
To stress-test its real-world reflexes, the developers brought in Tu Chenbo, a decorated regional champion and national second-tier badminton athlete, as a full-time sparring partner.
Within standard training shot distributions, the company reports XA1 hits over 99% of its returns, holding its own in high-speed rallies with competitive players.
That said, it is not invincible. Under high-intensity, continuous sessions, the robot still suffers occasional hardware freezes or algorithmic hesitations when faced with deceptive, high-velocity angles. The engineers are still busy revamping both mechanical durability and edge-case handling.
Why it matters
In embodied AI, locomotion and kinematic motion control—the so-called cerebellum—have matured rapidly. The true bottleneck remains the brain: real-time dynamic perception, strategic reasoning, and physical execution in non-stationary environments.
We think Shenchen Kaidong made a clever bet by targeting badminton. A court offers semi-structured boundaries, but a flying shuttlecock introduces brutal, millisecond-level non-linear aerodynamics. It turns a sport into a rigorous real-world proving ground for dynamic interaction.
Practically speaking, the immediate market is obvious. Elite training facilities and club courts could use XA1 as a tireless, pinpoint-accurate feeder machine, freeing coaches from mundane repetitive drills.
However, the massive catch right now is environmental overhead. XA1 currently leans on a fixed array of 30 external high-speed cameras watching the court. Until that bulky perception suite can be compressed onboard the chassis itself, widespread adoption will be heavily gated by installation costs.
Sources (1)
- zjnews.zjol.com.cn(opens in a new tab)Originalzjnews.zjol.com.cn