Self-Driving Cars Are Being Tested in a Whole New Way: By Blinding Them

Earlier this year, Stanford University's autonomous driving team, called the Intelligent Systems Laboratory, decided to test its self-driving cars in a more challenging and creative way. Instead of using cameras and sensors, the researchers equipped the cars with a monocular camera and a single lidar sensor, but with a twist: the camera was placed behind the car, capturing only a narrow field of view, while the lidar sensor was placed on the roof of the car, providing a 360-degree view.<br><br>This unique configuration was designed to simulate a scenario where the car is forced to rely on its perception and prediction abilities, rather than its sensors, to navigate through the environment. The researchers wanted to see how well the car could handle unexpected events, such as pedestrians stepping into the road or cars cutting into its lane, without the benefit of a wide-angle camera view or multiple sensors.<br><br>To test this setup, the researchers conducted a series of experiments on the Stanford campus, driving the car through various routes and scenarios, including narrow streets, intersections, and pedestrian crossings. They also simulated different weather conditions, such as fog, rain, and sunlight, to see how the car would perform in different environments.<br><br>The results of the experiments were impressive. Despite the limited field of view and lack of sensors, the car was able to navigate through the campus safely and efficiently, avoiding pedestrians and other obstacles with ease. The researchers attributed this success to the car's advanced perception and prediction algorithms, which were able to detect and respond to potential hazards even when the camera and lidar sensor were not providing a complete view of the environment.<br><br>This research has important implications for the development of self-driving cars, as it demonstrates the potential for these vehicles to operate safely and effectively even in challenging environments. By relying on advanced perception and prediction algorithms, rather than a multitude of sensors, self-driving cars may be able to navigate through complex scenarios with greater ease and accuracy. Additionally, this research highlights the importance of testing self-driving cars in a variety of scenarios, including those that simulate real-world challenges, in order to ensure their safety and reliability on the road.<br><br>Overall, the experiment conducted by the Stanford University researchers is a significant step forward in the development of self-driving cars, and it has the potential to shape the future of autonomous transportation.