AGIBOT A3's 10x faster vision enables human-level table tennis rallies
Using a 20kHz camera, AGIBOT A3 predicts ball trajectory in milliseconds for continuous rallies.
Developed with Peking University's SpikePingPong algorithm and Huang Tiejun's 20kHz high-frequency pulse camera, its vision response is 10x faster, enabling millimeter-level prediction and millisecond decisions for continuous rallies, trajectory tracking, whole-body planning, and seamless attack-defense switches. This level of high-speed dynamic control is exactly what real-world applications need: safer human-robot collaboration in factories, responsive service and elder care robots. A huge step forward from BAAI Conference 2026.
- Vision response is 10x faster thanks to a 20kHz high-frequency pulse camera and SpikePingPong algorithm.
- Enables millimeter-level trajectory prediction and millisecond decision-making for continuous table tennis rallies.
- Demonstrated at BAAI 2026 by AGIBOT and Peking University; technology applicable to factory safety and service robots.
Why It Matters
Proves high-speed dynamic control is viable for safer factories and more responsive service robots.