Tesla has started pushing a new Full Self-Driving Beta software update (10.5) to the fleet this weekend.
It includes some interesting new features based on the release notes, which you can find below.
Tesla is currently slowly rolling out what it is calling Full Self-Driving Beta (FSD Beta), which is an early version of its self-driving software 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 needs to remain vigilant and ready to take control at all times.
Since the responsibility lies with the driver and not Tesla’s system, it is still considered a level 2 driver-assist system despite its name. It has been sort of a “two steps forward, one step back” type of program as some updates have seen regressions in terms of the driving capabilities.
In recent updates, the automaker has been releasing more details about the changes to the driving behaviors in the release notes.
That’s again the case here for a new update (10.5) that Tesla started to push to owners in the Beta program in the US with a safety score of 98 and up.
Tesla wrote about the changes in version 10.5 in the release notes:
- Improved VRU (pedestrians, bicyclists, motorcycles) crossing velocity error by 20% from improved quality in our auto-labeling.
- Improved static world predictions (road lines, edges, and lane connectivity) by up to 13% using a new static world auto-labeler and adding 165K auto-labeled videos.
- Improved cone and sign detections by upreving the generalized static object network with 15K more video clips and adjusting oversampling and overweighting strategies (+4.5% precision, +10.4% recall).
- Improved cut-in detection network by 5.5% to help reduce false slowdowns.
- Enabled “emergency collision avoidance maneuvering” in shadow mode.
- Enabled behavior to lane change away from merges when safe to do so.
- Improved merge object detection recall by using multi-modal object prediction at intersections.
- Improved control for merges by increasing smoothness of arrival time constraints and considering possible merging objects beyond visibility.
- Improved lane changes by allowing larger deceleration limit in short-deadline situations.
- Improved lateral control for creeping forward to get more visibility.
- Improved modeling of road boundaries on high curvature roads for finer maneuvers.
- Improved logic to stay on-route and avoid unnecessary detours/rerouting.
Interestingly, Tesla provides actual data and percentage of improvement in certain capabilities when quantifiable.
Some notable points include Tesla improving static world detection by “up to 13%” by deploying a new auto labeling system that used 165,000 new videos.
Tesla has been relying heavily on a large team of thousands of “labelers” who literally review videos and mark what is visible to create clean data for Tesla’s machine learning systems.
But the automaker has also been developing technology to automatically label content for machine learning, and it sounds like it deployed a new one that helped train this new FSD version 10.5.
Another point that stood out to many people is the addition of “emergency collision avoidance maneuvering” in shadow mode.
It sounds like Tesla is adding the capacity for the vehicle to perform some more important emergency maneuvers to avoid accidents, which previously had more limitations due to Tesla limiting harsh steering maneuvers in its driver-assist systems.
The feature is currently in “shadow mode,” which means that the vehicles won’t actually perform the maneuver, but it will look for opportunities to do them based on real-world scenarios and send data to Tesla to see how it would react.
This new Full Self-Driving Beta 10.5 software is expected to be tested for the next few weeks before Tesla releases a new version and hopefully expands access to the software to more owners who paid for it.
The rollout to the fleet has so far been a lot slower than CEO Elon Musk initially announced.
Subscribe to Electrek on YouTube for exclusive videos and subscribe to the podcast.