Summary
- Full Self-Driving feature in Tesla vehicles was tested using a fake wall
- Former NASA engineer Mark Rober’s original video of a Model Y going through a painted wall sparked controversy
- Kyle Paul conducted a test using two Tesla EVs with mixed results, showing a older Model Y failed to detect the wall while a newer Cybertruck model successfully detected it
- Tesla vehicles use vision-only driver assistance system, sparking concerns about obstacle detection
- Latest hardware and software in Tesla vehicles have greatly improved obstacle-detection abilities
Article
Testing Tesla’s Full Self-Driving Software with Fake Wall: A Comparison
Introduction to the Experiment
Former NASA engineer and YouTuber Mark Rober recently conducted an experiment where a Tesla Model Y drove through a fake wall while the Full Self-Driving (FSD) software was supposedly active. The purpose of the experiment was to trick Tesla’s vision-only driver assistance system, which relies solely on cameras to sense surroundings. Many raised concerns about the objectivity of the test as Rober used basic Autopilot instead of FSD and utilized multiple takes.
Kyle Paul’s Follow-Up Experiment
Creator Kyle Paul conducted a follow-up experiment using two Tesla EVs to test the FSD software’s performance. The first attempt involved a Tesla Model Y equipped with the previous-generation Hardware 3 computer running FSD version 12.5.4.2. Despite the FSD being engaged, the Model Y failed to detect the fake wall and required manual intervention to avoid a collision. The test highlighted that older Tesla models may have lower FSD performance compared to newer cars.
Comparison Between Model Y and Cybertruck
After the Model Y’s failure to detect the obstacle, Paul tested the FSD software on a Tesla Cybertruck equipped with the latest Hardware 4 and FSD version 13 software. The Cybertruck showed a significant improvement in detecting the fake wall compared to the Model Y. The advanced cameras and processing capabilities of the HW4 hardware and updated software enabled the Cybertruck to come to a full stop without human intervention upon approaching the obstacle.
Real-World Implications and Discussion
While the scenario of driving through a fake wall is unlikely in real life, the experiments raise valid concerns about Tesla’s reliance on video cameras for driver aids. The reluctance to incorporate radar and LiDAR technologies may impact the accuracy and reliability of Tesla’s FSD software. However, the results from the Cybertruck test suggest that Tesla has made significant improvements to its hardware and software, enhancing obstacle detection capabilities.
Future Expectations for Older Tesla Models
It is possible that older Tesla models with Hardware 3 may receive updates to FSD version 13 in the future, potentially improving their obstacle detection abilities. However, the performance gap between older and newer Tesla vehicles highlights the importance of regular software updates and advancements in hardware technology. Customers driving older Tesla models may expect enhancements in FSD functionality through future updates.
Conclusion
In conclusion, the experiments conducted by Mark Rober and Kyle Paul shed light on the performance disparities of Tesla’s FSD software on different hardware generations. While older models like the Model Y with Hardware 3 may struggle with obstacle detection, newer vehicles like the Cybertruck equipped with Hardware 4 demonstrate significant advancements in FSD capabilities. As Tesla continues to innovate and improve its driver assistance systems, customers can anticipate enhanced safety features and performance in future software updates.
Read the full article here
