Tesla's autonomous guide deceived by a painted wall? The test that makes you discuss

Tesla’s autonomous guide deceived by a painted wall? The test that makes you discuss

By Dr. Kyle Muller

The video of a YouTuber would demonstrate how the Tesla Autopilot system can be deceived by a “wall” painted in order to reproduce the appearance of a road …

The system of Autonomous Autopilot Guide of Tesla It relies exclusively to the cameras to detect obstacles and follow paths. But how much it is reliable Really? This is what the ex-engineer NASA and YouTuber Mark Rober asked himself, who decided to test him: according to what is clear from his video (see below), would have painted a wall with the image of a free road in front, In perfect Willy style the Coyote, managing to trick The car that, in “autonomous guide” mode, crashed without even mentioning a braking.

Vision problems?

Tesla has anal eliminated sensors like radar (system that uses radio waves to detect the distance of objects) e Lidar (who uses laser impulses to create three -dimensional maps of the surrounding environments), defining them – according to Elon Musk – “stupid, expensive and useless”. However, according to Rober’s test, The only cameras may not be enough to avoid ambiguous situations or visual illusions.

The disturbing result (but to be verified)

In his experiment Rober compared the Tesla Autopilot with a Lexus car equipped with Lidar (provided by the Luminar company). The lexus would have brilliantly passed all the testsincluding those with rain And fogWhile The Tesla, as you can see in the video, overwhelmed a mannequin that simulated a child and ignored the obstacle when immersed in one fog artificial.

Cartoon scenario, real consequences

The published video would highlight one significant vulnerability in the systemjust as Tesla’s number one, Elon Musk, announces for this year a version of the “non-supervised” full self-driveing ​​(ie without necessary human supervisor) and a future service of robotaxi.

In short, if all you see from the video is not the result of manipulations or alterations (and to say the truth there is no obvious clue that leads to thinking it) this would raise serious questions about the safety of autonomous cars based exclusively on artificial vision.

Kyle Muller
About the author
Dr. Kyle Muller
Dr. Kyle Mueller is a Research Analyst at the Harris County Juvenile Probation Department in Houston, Texas. He earned his Ph.D. in Criminal Justice from Texas State University in 2019, where his dissertation was supervised by Dr. Scott Bowman. Dr. Mueller's research focuses on juvenile justice policies and evidence-based interventions aimed at reducing recidivism among youth offenders. His work has been instrumental in shaping data-driven strategies within the juvenile justice system, emphasizing rehabilitation and community engagement.
Published in