Tuesday, February 21, 2017

Tesla is Using AI and Big Data to Drive Autonomous Vehicles forward

(Source: Tesla Motors)

Autonomous vehicles, better know as self-driving cars, require a lot of detailed data about the roads on which they travel.  Almost every automobile manufacturer is developing autonomous vehicles, including BMW, Ford, GM, Mercedes and Tesla.  However, unlike other manufacturers, Tesla is leveraging the cars it already has on the road to continuously improve the function of all of its vehicles.  Every Tesla sold is equipped with self-driving hardware and software, referred to by the company as Autopilot.  The cars come equipped with 12 sensors on the bottom or the car, a front-facing camera and radar.  All Teslas are also wirelessly connected to the cloud and continuously upload data to the company.  Tesla can collect data from its vehicles, regardless of whether they are operating in Autopilot mode or not.  Thus far, Tesla has collected more than 1.3 billion miles of real world data from its cars and adds approximately 1.5 million miles of data each day.  By way of comparison, to date the Google self driving program has only acquired 2 million miles of data.  This use of Big Data gives Tesla a significant advantage over their competitors.  One such example are the road maps these vehicles must use.  By incorporating data from their fleet, some Tesla maps are now 100 times more detailed than commercially available GPS maps.
   Tesla also uses their huge data advantage to continuously improve the operation of their vehicles.  Every car is equipped with machine learning software, a type of artificial intelligence.  As stated earlier, the cars are always connected to the cloud and to each other wirelessly.  As Elon Musk, the Founder and CEO of Tesla puts it, "The whole Tesla fleet operates as a network.  When one car learns something, they all learn it."  This continual learning approach, paired with their huge data advantage, puts Tesla in a strong position to pull away from their competitors in the race for self-driving vehicles.


  1. Great and informative post Doug! It's great to see Tesla advancing the auto/techno space (my language) forward. I pray that in their quest to move rapidly they are paying close attention to the pitfalls of rapid technology advancement in automobiles. For example, didn't a few of the Tesla Model S vehicles catch fire. I'm not sure the public received a clear explanation of what went wrong. The fact is rapid advances in automobile technology come with a cost. That of course needs to be mitigated against potential lives that can be negatively impacted. I think it's fair to say there are still questions that need to be answered regarding the safety of the Tesla Model S.

    Here's some alarming data I just pulled up.....U.S. fire departments responded to an estimated average of 152,300 automobile fires per year in the 2006-2010 time period. These fires caused an average of 209 civilian deaths, 764 civilian injuries, and $536 million in direct property damage.

    While this shouldn't be solely blamed on car companies seeking to enhance the driving experience by doing things in a rapidly technological timeframe, safety must continue to be their numero uno priority. I trust Elon Musk and Tesla have that at the top of their list as they seek to become the numero uno car manufacturer.

  2. Anthony,

    Thanks for your comment. You're absolutely right, there is always a balance between technological advancement and safety. As consumers of technology, we depend on the companies and regulators to get this balance right. I didn't know that there were so many automobile fires and fire-related deaths each year. Thanks for the info!


  3. I'm really skeptical of the future of self-driving cars. When road conditions are ideal, I'm sure they work fine, but what about when they aren't? Road construction, road closures, poorly marked lanes, lane closures. How the heck is a computer supposed to know which side of the haphazardly placed orange cones to drive on? Sometimes I can't even tell!

    And don't forget about jaywalkers. And those drivers who fly through the yellow/red light hoping to get through the intersection before cross-traffic starts moving. Or those difficult left turns where traffic is so heavy that the only way to make the turn is to do it once the light turns red and oncoming traffic stops (LA, I'm looking at you). So many edge cases, so many chances for problems. I just don't see a computer figuring this stuff out. Prove me wrong, Tesla and Google, prove me wrong!

  4. You may well be right. I think that unless there are significant issues with the technology, autonomous vehicles will be introduced. However, I'm not planning to be an early adopter.