Why are real-time vision and Edge AI a bottleneck on open pitches?

Answer

A robot must decide in fractions of a second, so it cannot wait for a round trip to the cloud — processing happens locally on the device (Edge AI) under tight compute and power budgets. And the wider the pitch, the harder depth perception and ball tracking become: a distant ball is a handful of pixels, and estimating its distance and speed from a 2D image grows less accurate with range. Open space multiplies the computational load precisely when resources are scarcest.