Event-Triggered Vision: How Sony's Table Tennis Robot Cuts Processing Overhead

Sony AI's table tennis robot Ace uses event-based vision sensors instead of standard high-speed cameras to track balls at 30+ meters per second. These sensors respond only to brightness changes, eliminating blur and drastically reducing data volume compared to traditional cameras capturing 200–1000 frames per second. The system maintains microsecond precision while cutting processing demands, pointing toward a practical approach for time-critical robotics applications.