In late 2020, Tesla released the next generation of the company’s full self-driving software called FSD Beta.
This software version is a significant architectural rewrite of Tesla’s Autopilot software and unlike previous iterations of Autopilot, requires merging all images coming from 8 cameras into a vector space that the vehicle uses to execute driving in the real world.
In simple terms, the new software creates a representation of the real world by piecing together images coming from the 8 cameras, and the vehicle uses this virtual representation of the real world to make driving decisions.
The new approach has allowed Tesla to enable vehicles to tackle more complex driving settings such as surface streets, roundabouts, and unprotected left turns, which are currently impossible for Tesla’s previous Autopilot software.
Tesla initially released FSD Beta to the 3000 or so individuals in the company’s Early Access Program. However, that number has currently grown to more than 160,000 vehicles and Elon Musk has recently said by the end of this year FSD Beta will be running on 1 million Teslas.
Since FSD Beta’s initial rollout in late 2020, the software has seen several upgrades which have improved the smoothness and capabilities of the self-driving software.
Tesla’s last FSD update was FSD Beta 10.69. This iteration of FSD Beta was released at the end of August and has brought with it several improvements getting Tesla one step closer to the goal of having the software completely take over the driving task.
Having said that, FSD Beta 10.69 is still far from a level 5 fully autonomous solution and the software still needs a lot of work to make it fully self-driving.
One significant area the current iteration of FSD Beta could improve is on the highways. This improvement has less to do with functionality but rather with the fundamental software architecture the vehicle uses to navigate highways.
On highways Tesla still uses the pre-FSD Beta iteration of autopilot that doesn’t use all 8 cameras to create a vector space and heavily relies on manually written codes as opposed to AI to make driving decisions. However, the Tesla autopilot team has been working hard to bring FSD Beta to the highway.
And on November 11, all that hard work has paid off and Tesla has started releasing FSD Beta V11 to its fleet of vehicles. Tesla has been cautious and has only rolled out this new software version and as a result there isn’t a lot known about the software.
However, today we have gotten more clarity regarding all the improvements coming with FSD Beta V11 as the release note for the new software have shared online. And the release note lists all the improvements Tesla has made over the last couple of months. And some of the improvements Tesla lists are…
– Enabled FSD Beta on the highway. This unifies the vision and planning stack on and off-highway and replaces the legacy highway stack, which is over four years old. The legacy highway stack still relies on several single-camera and single-frame networks and was set up to handle simple lane-specific maneuvers. FSD Beta’s multi-camera video networks and next-gen planner, which allows for more complex agent interactions with less reliance on lanes, make way for adding more intelligent behaviors, smoother control, and better decision-making.
– Improved Occupancy Network’s recall for close-by obstacles and precision in severe weather conditions with a 4x increase in transformer spatial resolution, 20% increase in image featurizer capacity, improved side camera calibration, and 260k more video training clips (real-world and simulation).
– Improved merging behavior by leveraging lane geometry and lane bounds, association with coarse map information, and better gap selection algorithms, allowing for a smoother and safer experience.
– Added highway behavior to offset away from blocked lanes and generic obstacles like road debris while also adding a smooth hand-off between in-lane offsetting and lane changing.
– Improved speed-based lane change decisions to better avoid slowing down traffic in fast lanes, and interfere less with navigation.
– Reduced sensitivity for speed-based lane changes in CHILL mode.
– Improved lane changes to allow higher jerk maneuvers if required to stay on route or to move away from lane blockages.
– Improved smoothness at highway lane splits by being less strict about centering between lane lines and allowing lower jerk maneuvers, where safe to do so.
– Reduced latency of trajectory optimization by 20% on average, without sacrificing behavior, by leveraging numerical tricks for more efficient computations.
All the improvements Tesla listed on the FSD Beta V11 release note are certainly exciting, but, since the software update is just rolling out, we have yet to see the software in action in the real world.
However, we will be sure to keep you posted once FSD Beta V11 drives start being shared. Until then, make sure to visit our site torquenews.com/Tesla regularly for the latest updates.
So what do you think? Excited about all the improvements coming with FSD Beta V11? Also, how long do you think it will take Tesla to make FSD a fully self-driving system? Let me know your thoughts in the comments below.
Image: Courtesy of Tesla, Inc.
For more information check out: Tesla Demand has "Fallen Off a Cliff" -Kelly Blue Book
Tinsae Aregay has been following Tesla and The evolution of the EV space on a daily basis for several years. He covers everything about Tesla from the cars to Elon Musk, the energy business, and autonomy. Follow Tinsae on Twitter at @TinsaeAregay for daily Tesla news.