Tesla releases new beta update for full self-driving with improved object recognition


Tesla has started to bring a new Full Self-Driving Beta software update (version 10.6) with improved object recognition to its fleet.

It’s the latest step on the long road to improving the FSD package and delivering on a multi-delayed customer promise.

Tesla is currently slowly rolling out what is known as the Full Self-Driving Beta (FSD Beta), an early version of its self-driving software that is currently being tested by a fleet of Tesla owners selected by the company and through its “safety test”. Score. “

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

Since the responsibility rests with the driver and not with Tesla’s system, despite its name it is still considered a level 2 driver assistance system. It was sort of a “two step forward, one step back” program as some updates saw regression in terms of driving skills.

In the last few updates, the automaker published more details about the changes in driving behavior in the release notes.

Last month we reported on the FSD 10.5 update in the release notes.

Tesla begins today to offer US owners a new version of the Full Self-Driving Beta with a safety rating of 98 and above.

Tesla wrote about the changes in version 10.6 in the release notes:

  • Improved network architecture for object recognition for non-VRUs (e.g. cars, trucks, buses). 7% higher memory. 16% lower depth error and 21% lower speed error for crossing vehicles.
  • New visibility network with 18.5% less mean relative error.
  • New general static object network with 17% precision improvements in cases with high curvature and night.
  • Improved stopping position on unprotected left turns while giving way to oncoming objects, using object predictions beyond the intersection.
  • Leave more room for lengthwise alignment during the merge by including modeling the end of the merge area.
  • Improved convenience when moving objects that cut out of your lane.

The automaker has stated that the rate of improvement will increase with its Dojo supercomputer.

Earlier this week, we reported that Tesla released new footage of its auto-labeling tool for its self-driving efforts.

With the combination of auto-labeling and dojo, we could see this accelerated rate of improvement.

FTC: We Use Income Earning Auto Affiliate Links. More.


Subscribe to Electrek on YouTube for exclusive videos and subscribe to the podcast.

About Willie Ash

Check Also

The new PS5 system software beta adds two of the most requested features by users

A folder of sorts finally hits the PS5 system software later this year. Sony Native …