Elon Musk, CEO Tesla, menekankan visi masa depan perusahaan yang fokus bukan hanya pada kendaraan listrik, tetapi juga pada teknologi self-driving. Tesla’s approach to self-driving technology has been unique, relying solely on cameras and artificial intelligence rather than additional sensors like radar and lidar. A recent video uploaded by YouTuber and engineer Mark Rober highlights a test that demonstrates the limitations of this approach. The video features an experiment where a Tesla on Autopilot is tricked by a painted “fake wall” to see if it will crash into it. While Tesla’s engineers have not yet found a way for their vehicles to navigate through fake tunnels like in Looney Tunes, the test showed the car crashing into the wall, much like Wile E. Coyote.
In comparison, Rober tested a Lexus SUV equipped with lidar technology from Luminar. Both vehicles initially passed a test where they stopped for a simulated child crossing the road. However, the real difference emerged when Rober and his team used a painted wall to create a false environment. The Lexus, with lidar, detected the wall and stopped, while the Tesla did not. This experiment raised concerns about the reliance on camera-based systems in Tesla’s self-driving technology, especially when it comes to safety and precision in identifying obstacles. Despite Tesla’s advancements in Autopilot and Full Self-Driving features, concerns persist over the system’s performance and its involvement in accidents. Experts suggest that Tesla’s camera-based approach, while similar to human driving, may not be sufficient to create a truly safe and reliable autonomous vehicle. The video serves as a reminder of the potential limitations of Tesla’s technology and prompts questions about the future of self-driving cars.
