Tesla Implementing Mobileye’s Autopilot Quicker Than Anyone Else
Originally published on EV Obsession.
The recent Quarter 3 conference call held by Mobileye yielded some interesting new information, including some interesting comments about the company’s relationship with Tesla.
On that count, it was apparently stated (tip of the hat here to “Cosmacelf” on the Tesla Motors Club forum for covering the call) that Tesla is essentially unique amongst auto manufacturers for its willingness to push the company’s system faster, and truly to its limits.
Also very noteworthy was the comment that “True Autopilot” would begin its (slow) rollout next year — and that it will require the integration of an 8-camera suite, radar + sonar systems, and 5 EyeQ3 chips. The rollout of autopilot software will apparently take several years to be completed.
Here’s more from the recent conference call, via Cosmacelf:
Mobileye has 2 production agreements for semi-autonomous driving for 2016 (presumably one is Tesla), another manufacturer coming on line in 2017 and one more in 2018.
Tesla’s current self learning capability is Tesla specific. Mobileye is not involved in this at all. Mobileye gets its training data when OEMs do validation testing. I guess the big automakers do a lot of validation testing and Mobileye thinks this is sufficient for their purposes (they pretty much dismissed the Tesla fleet info and said that the validation data set was huge).
My comments based on this and Elon’s Autopilot conference call: It is clear that Tesla isn’t just taking what Mobileye is giving them. According to Elon, they have augmented the base Mobileye system with high resolution GPS maps, driver assisted learning, fleet learning, and rapid updates. The high resolution GPS maps, along with driver input, allows the car to behave correctly in ambiguous situations. If Mobileye technology is confused about what is going on in a particular area, it will always be confused in that area (barring very slowly changing model updates which may come out once a year (my estimate)). So Tesla tracks what the human does in that area, averaged over the fleet. If the Mobileye model says to follow the right lane marker at a certain GPS co-ordinate and exit the freeway, while the driver grabs the wheel and keeps the car on the freeway, then the Tesla software learns that behavior for next time and additionally uploads that info for fleet learning. According to Elon, fleet learning updates could occur weekly (presumably more often in the heavier traveled areas).
Certainly an interesting subject to speculate on. I’m very curious myself to see how Tesla’s autonomous system compares to the Mobileye systems put into use by other automakers over the coming years. As noted above, Tesla does seem to have covered all of the bases and come up with a good plan of action to deliver a particularly effective autonomous system….
On that note, it’s probably worth mentioning here that the Quarter 3 conference call also revealed that Mobileye’s Q3 2015 revenue ($70.6 million) was 104% higher than Q3 2014’s ($34.7 million).
Have a tip for CleanTechnica? Want to advertise? Want to suggest a guest for our CleanTech Talk podcast? Contact us here.
CleanTechnica Holiday Wish Book
Our Latest EVObsession Video
CleanTechnica uses affiliate links. See our policy here.
Regular cruse control can get you in just as much trouble between falling asleep at the wheel or texting while driving and crashing into someone. Hard to regulate careless. Seems like a great feature when used sanely.
I just posted a video on Youtube with me using cruze kontrol. It’s crazy, man!
Got a link? Want to describe your experience?
I tried to post the link but was denied by the site. Let’s see if I can sneak it past.
It’s doubleUdoubleUdoubleU. sarcasmdotcom
Put some explanation of why you’re posting the link along with the link.
I’m very short on time right now and don’t have time to open each naked link to see what it is and open the conversation to see if it’s relevant.
We get hit with a lot of spam and a fair amount junk. Waists a lot of my time.
Cruz control would be pretty useful for Jeb Bush’s campaign right about now…
TSFS (That’s some funny sh!!) hahahaha
hahaha 😀
We now have more than 8,000 miles of Autosteer experience in our Tesla. Although I have noticed no “learning” updates so far, I think it will surely be effective; in the “confused” situations I have encountered, averaging out the reaction of just a few Tesla drivers will likely eliminate any concerns. Autosteer is a really a cool – and relaxing – way to “get there!”
A recent trip report is posted here:
http://www.teslatouring.com/ehea2-a.html#sthash.BxyZdpFn.dpbs
or here:
http://www.teslamotorsclub.com/showthread.php/56457-Yorba-Linda-Vacationers-First-To-Go-Coast-To-Coast-Mostly-by-Autosteer
What if there’s temporary construction and cones directing drivers to move over into one of the oncoming traffic lanes? I hope Tesla fleet learning doesn’t direct drivers into oncoming traffic after the cones are removed… I’m sure it wouldn’t, but it’s an interesting problem with fleet learning of temporary detours. It’s still hard to believe the system could ever account for every possibility… Maybe with enough time and learning from mistakes, but unfortunately every mistake could lead to driver fatalities so the whole system makes me nervous.
Basic Rules Number One and Two
1) Don’t run into stuff
2) Don’t drive over cliffs
If the car can’t determine the appropriate route around a detour or “unlearn” the detour path then it should simply park itself. Driving into oncoming traffic would be prevented by Rule One.
Here’s what I see further down the road. Self-driving pods with no steeringwheels like Google is working with.
Those very few times a car parks itself out of inadequate data to proceed, driving will be taken over by a “drone operator” and the system will be quickly updated/patched.
Ha, interesting, and yeah, this is not my area of expertise, but it seems like the system would “learn” a bad habit.
It looks like Tesla is learning, and advancing beyond their vendor/supplier… yet again.
The Tesla Roadster originally had a drive train sourced by AC Propulsion… but Tesla moved beyond them.
Yes, there’s also word that Tesla is adding on to what Mobileye provides (in the software). Or was that in the article above… anyhow, yes, it just seems more eager to move faster than anyone else on both EVs and autopilot. And is throwing the best engineers it can find at that.
Google gives you a great Chance to make 18648$/2weeks at your home.If you are some intelligent you make many more.I am also earning many more, my relatives wondered to see how i settle my Life in few days thank GOD to you for this… You can also make cash i never tell alie you should check this I am sure you shocked to see this amazing offer… I’m Loving it!!! ☻ ▼ ▼ ▼
kmt…….
✔✔✔✔✔✔✔➤➤ http://www.dailycareeersplans.tk
.
░▒▓█▀▄▀▄▀▄█▓▒░░▒▓█▀▄▀▄▀▄█▓▒░░▒▓█▀▄▀▄▀▄█▓▒░
Didn’t Tesla have to bring the drivetrain inhouse when the outsourced units couldn’t stand up to the torque?