The Stigmergic Adaptive Responsive Learning Crossing prototype — also known as the Starling Crossing — uses LEDs and an army of cameras to illuminate the road based on the movements of pedestrians, bicycles and cars. It works by using a neural network to analyze the feed from the cameras and differentiate between cars, bikes and pedestrians, and then anticipate the movements of each. Then the surface of the road, which is capable of bearing the weight of multiple vehicles and maintains traction when wet, illuminates to guide the different users and tell them when to stop, slow down and go.

The team behind this tech, called Umbrellium, is composed of architects, designers, tactical urbanists and creative technologists all with a common goal in mind — improving pedestiran safety. For the time being, this is only a temporary design study in South London, and there is yet no mention of a real-world rollout.

See the Starling Crossing in action here.