Beihang's new research has been featured in "Nature Communications", enabling robots to have human-level visual capabilities.

On February 10, a research team from Beihang University published a study in Nature Communications, stating that their developed visual system processes motion four times faster than current mainstream optical flow technology, promising to significantly improve the response speed of systems such as autonomous vehicles, industrial robots, and surgical robots. The researchers developed a neural morphology module that can directly detect changes in light intensity over time in a region, allocating computing power only to the parts that change, achieving real-time motion recognition. In simulated driving and robot operation tests, the system reduced latency by about 75% and improved tracking accuracy by double in complex motion scenarios.