Tesla's Latest Autopilot Update Is A Hit With Owners
Tesla (NASDAQ: TSLA) owners are gushing about the latest version of Autopilot for Hardware 2 cars. One Tesla owner said to me that, with the new update, Autopilot "feels like a human driver." On the Tesla Motors Club forum, an owner going by MikeH said that Autopilot is "now as good or better than" him at changing lanes on the highway.
From article, (Tesla (NASDAQ:TSLA) owners are gushing about the latest version of Autopilot for Hardware 2 cars. One Tesla owner said to me that, with the new update, Autopilot “feels like a human driver.” On the Tesla Motors Club forum, an owner going by MikeH said that Autopilot is “now as good or better than” him at changing lanes on the highway. Dozens of comments on the forum are similarly effusive.
The update brings three major improvements:
- greatly improved lane keeping: a car now recognizes the lane that it’s in more reliably, and holds steady in the centre of the lane instead of “ping-ponging” between lane lines
- sharp curves and hills that previously confounded Autopilot are now handled with grace
- greatly improved automatic lane changes on the highway
Over the past five months, Tesla has indicated that is has been reworking its Autopilot software stack at a fundamental level. In its Q3 2017 update letter, Tesla wrote:
Now that the foundation of the Tesla vision neural net is right, which was an exceptionally difficult problem, as it must fit into far less computing power than is typically used, we expect a rapid rollout of additional functionality over the next several months and are progressing rapidly towards our goal of a coast-to-coast drive with no one touching the controls.
This comment suggests that Tesla changed to a new neural network architecture as recently as Q3 or Q4 2017. In its Q4 2017 update letter, Tesla expanded on this theme:
For More InfoThe upcoming autonomous coast-to-coast drive will showcase a major leap forward for our self-driving technology. Additionally, an extensive overhaul of the underlying architecture of our software has now been completed, which has enabled a step-change improvement in the collection and analysis of data and fundamentally enhanced its machine learning capabilities. Our neural net, which expands as our customer fleet grows, is able to collect and analyze more high-quality data than ever before, enabling us to rollout a series of new Autopilot features in 2018 and beyond.)
No comments:
Post a Comment