May 22, 2022


Complete Australian News World

Tesla begins releasing a major beta update for fully autonomous driving

Tesla begins releasing a major beta update for fully autonomous driving

Tesla has started releasing a big new update to its Full Autonomous Driving (FSD) Beta software with lots of improvements.

The automaker may expand access to the beta version if this update “works well,” according to CEO Elon Musk.

A fully self-driving Tesla Beta

Since October 2020, Tesla has begun rolling out what it calls a “Full Self-Driving Beta” (FSD Beta), an early version of the self-driving software currently being tested by a fleet of Tesla owners. Chosen by the company through “Safety test score. “

The software enables the vehicle to drive autonomously to a destination entered into the vehicle’s navigation system, but the driver needs to remain alert and ready to be in control at all times.

Since the responsibility lies with the driver and not the Tesla system, it is still considered a Level 2 driver assistance system despite its name. It was a kind of “two steps forward, one step back” type of program, with some updates seeing a regression in terms of driving capabilities.

Tesla frequently releases new software updates for the FSD Beta and adds more owners to it.

The last important update was FSD Beta 10.10.1 Update In early February.

As of the fourth quarter of 2021, the automaker said it had nearly 60,000 owners in the FSD Beta program.

Tesla FSD Beta 10.11.0 Update

Tesla has now started pushing out a new FSD Beta 10.11 early access software update, which is significant based on the release notes.

Here are the release notes:

  • Improved lane geometry modeling from a dense raster (“bag of points”) to a self-regressive decoder that predicts and directly connects point-to-point ‘vector space’ lanes using a neural network of switches. This enables us to predict cross-lanes, allows for computationally cheaper and less error-prone subsequent processing, and paves the way for predicting many other signals and their relationships jointly and end-to-end. Use more accurate predictions of where vehicles will turn or merge to reduce unnecessary deceleration for vehicles that will not cross our path.
  • Improve understanding of the right of way if the map is inaccurate or the vehicle cannot follow navigation. In particular, modeling intersection ranges are now entirely based on network predictions and no longer use map-based inference methods.
  • The accuracy of VRU detections is improved by 44.9%, which significantly reduces pedestrians and bicycles false positives (especially around tar layers, slip marks and raindrops). This was achieved by increasing the data volume of the next generation of auto-tagging, training previously frozen network parameters, and modifying network loss functions. We find that this reduces the incidence of VRU-related pseudohysteresis.
  • The expected speed error for motorcycles, scooters, wheelchairs and pedestrians too close was reduced by 63.6%. To do this, we presented a new data set to simulate high-speed VRU interactions for antagonism. This update improves the autopilot control around fast-moving and cutting VRUs.
  • Improved crawler profile with higher jerk at crawl start and finish.
  • The control of close obstacles is improved by predicting the continuous distance of the static geometry using the general static obstacle network.
  • The vehicle “stopped” attribute error rate was reduced by 17%, by increasing the dataset size by 14%. Brake light accuracy has also been improved.
  • The Clear-to-Go scenario speed error has been improved by 5% and the Highway scenario speed error by 10%, which is achieved by adjusting the loss function aimed at improving performance in challenging scenarios.
  • Improved detection and control of open car doors.
  • Improved smoothness through cornering using an optimization-based approach to identify road lines unrelated to control due to lateral and longitudinal acceleration and vibration limits as well as vehicle kinematics.
  • The stability of FSD Ul visualizations is improved by optimizing the Ethernet data transmission pipeline by 15%.
READ  Hostage situation ends at Apple Store Amsterdam | news | DW

CEO Elon Musk says that if this update “works well,” Tesla will “likely” reduce access to drivers with a “95” driver safety score. This would give a lot of FSD buyers access to the trial version.

As for Canada, Tesla owners who bought FSD have been able to hit the driver safety score for more than a week now. The automaker is expected to soon begin releasing the beta for those with the highest scores.

FTC: We use affiliate links to earn income. more.

Subscribe to Electrek on YouTube for exclusive videos and subscribe to audio notation.