This demonstrates how LINGO-1 reasons about potential hazards while driving at night. You can see the model first verifies that the scenario involves driving in the dark and acknowledges the need to be cautious of the following potential hazards: [A] a narrow drivable area, [B] a cyclist on the road, and [C] the presence of a car in front, which is challenging to spot in this image. It also correctly recognises that “the cyclist could make sudden moves or turns, so it’s important to keep a safe distance and be prepared to react accordingly.”

(video) Ride The Wayve: Autonomous driving in different weather conditions

GIPHY App Key not set. Please check settings