Tesla has issued yet another giant update for its Full Self Driving (FSD) software, dramatically improving vehicle behavior in numerous situations.
Despite the controversy that seems to consistently surround Tesla’s autonomous driving suite, it just keeps improving, with new features and capabilities added at an astounding rate. Now, Tesla has issued yet another dramatic update to Full Self Driving, bringing significant improvements in vehicle intelligence and the dream of fully autonomous vehicles one step closer to reality.
@Winnersechelon first reported the newest FSD update on Twitter, posting a screenshot of the release notes for the update.
Release notes for latest FSD Beta release pic.twitter.com/0lDmZbmEUb
— TeslaBull FSD BETA (@Winnersechelon) April 17, 2023
Perhaps more than previous updates, 11.4 focuses on vastly improving vehicle behavior, with new features taking a back seat. According to Tesla’s release notes, those improvements are seen throughout numerous environments.
Foremost, Tesla FSD is better than ever at recognizing and reacting to pedestrians that are either about to cross the street or are already doing so. This stems from an improvement in kinematic calculations, which allows the vehicle to predict where people, vehicles, and objects are going. Regarding pedestrians, the vehicle will act more smoothly, either progressing through the intersection or stopping.
This improvement also extends to other “vulnerable road users,” such as bikers and cyclists, which Tesla FSD will now react more quickly to and slow down for if necessary.
Another improvement that urbanites can enjoy is in “turn performance.” When in dense city areas or unexpected turns, the vehicle will now follow the turn more correctly, dodging parked vehicles and avoiding bus lanes. This improvement is paired with an update to bus lane recognition overall, helping to prevent the vehicle from being where it isn’t supposed to be.
Other updates focus on improving performance in more rural areas. One such improvement is improved lane, line, and edge of road detection, which can be helpful in areas with less consistent road painting and paving. This update feeds into two others, improved lane guidance, ensuring the vehicle is in the correct lane to reach its destination, and improved prediction of the movement of “partial cut-ins,” or cars that haven’t wholly dedicated to a lane. Tesla’s massive library of clips and auto-labeling system makes both upgrades possible.
In the case of improved lane guidance, Tesla states that this update will reduce driver interventions by 64%, ensuring the vehicle isn’t in the wrong lane and reaches the destination more consistently.
Another update affecting how Tesla vehicles behave in lanes is lane change speed control. Tesla FSD will now consider “upcoming navigation deadlines” and will use more or less speed if necessary to switch lanes. Moreover, the vehicle will more accurately calculate the number of lanes it needs to change, thanks to Tesla’s ever-improving vehicle network.
Vehicle speed also improves in countless other situations as well. Thanks to Tesla’s all-new “Vision Speed” network, its network of vehicles can now infer the typical driving speed of any given roadway, improving how the vehicle reacts to areas that typically require lower speeds, such as neighborhoods or parking lots.
The final speed update regards how the vehicle reacts to the weather. The max autopilot speed will now take into account current weather conditions and limit driving speed if visibility or traction is too poor. According to Tesla’s release notes, this calculation is quite advanced, as it will take into account not only its condition, such as tire tread, and its data on visibility and road wetness, but it will also account for outside factors, such as tire spray from other vehicles and traffic intensity.
The final two updates are far more general and will improve the vehicle’s autonomous operation overall. First, Tesla has improved “long-range path blockage,” meaning that its vehicles will more quickly merge into other lanes to avoid obstacles in the road. Second, Tesla has reduced “photon-to-control” latency by 2%, meaning the vehicle will react 2% faster to anything it “sees.”
What do you think of the article? Do you have any comments, questions, or concerns? Shoot me an email at william@teslarati.com. You can also reach me on Twitter @WilliamWritin. If you have news tips, email us at tips@teslarati.com!
