Recently, Musk stated on overseas social platforms that Tesla has begun to push the FSD Beta V11 version of the assisted driving system to testers, and overseas users will receive the push starting from 15:11 on the 12th. Musk also confirmed that Tesla will be "widely rolled out" to more North American users in a few weeks.
Judging from the release notes, FSD Beta V11 (2022.40.5) has been mainly updated and optimized in the following aspects:
Enable FSD Beta on highways. FSD Beta’s multi-camera video network and next-generation planning system allow for more complex agent interactions with fewer lane dependencies, making way for smarter behavior, smoother control and better decision-making.
Improved the accuracy of the Occupancy Network in close-range obstacle recall and adverse weather conditions, improved the spatial resolution of the transformer by 4 times, increased the capacity of the image feature by 20%, and improved the side camera calibration. And added 260,000 video training clips (including actual and simulation).
Reduced the prediction speed error of motorcycles, scooters, wheelchairs and pedestrians at very close distances by 63.6%.
Improved crawling profile to increase acceleration at start and end.
Improved control of nearby obstacles by using a general static obstacle network to predict continuous distances from static geometry.
By increasing the dataset size by 14%, the vehicle "parked" attribute error rate was reduced by 17%, while also improving the accuracy of brake lights.
The performance of difficult scenes was improved by adjusting the loss function, improving the speed error of clear scenes by 5% and the speed error of highway scenes by 10%.
Improved detection and control of opening car doors.
When calculating vehicle kinematics, the smoothness of turns is improved by using optimization-based methods to decide which road lines are independent of the calculation of given lateral and longitudinal accelerations and steep lift limits.
Optimize the Ethernet data transmission pipeline by 15%, thereby improving the stability of FSD UI visualization.
Improved the recall rate of vehicles behind the ego and improved the accuracy of the vehicle detection network.
According to Bitauto App’s “Hot List” data, Tesla (imported) ranks 89th among all brands. If you need more data, please check out Bitauto App.
Editor: Zhu Yuxiang