Its Robostral Navigate AI model needs input from just one color camera, doing without Lidar, depth sensors, or multiple ...