Why Tesla won't use LIDAR for full self driving cars
Alex Sibila Alex Sibila
31.2K subscribers
206,081 views
0

 Published On Apr 23, 2019

Just about every company working on Autonomous vehicles is using lidar as its main way of detecting obstacles. Self-driving projects from Google, Uber, and Toyota all stick a sensor on top that helps it navigate the world. Which makes sense, Lidar has a lot of advantages over cameras and even radar. The technology works by transmitting lasers and then measuring the reflections to determine how far away things are, but with a lot of accuracy. You essentially get a 3D mapped version of the world using this technology, which you would think has a lot of advantages when you’re trying to create something that can maneuver itself through the world. But Tesla thinks otherwise - Elon made it very clear on Autonomy Day that Tesla would not be using lidar in there cars. But why? Well there’s a couple reasons.

First off, its really expensive. Even though the cost of Lidar has dropped in recent years, Google’s subsidiary Waymo still pays around $7500 for there lidar sensors - not something you want to tack on to the price of an already expensive car. Not to mention repair costs if it were to break or malfunction.

Secondly, and probably Tesla’s main selling point for just using a camera and radar setup - is that the road is designed for people that can see. Every sign, every road marking, every intersection is full of visual stimuli to tell you what to do with your car. Not to mention obstacles, a camera and radar using Tesla’s system is able to tell that there is an obstacle in the way and even tell what type of obstacle it is - something lidar alone cannot do. The example they gave was a plastic bag rolling across the highway. Lidar alone will see this obstacle and slam on the brakes - something you definitely don’t want to do when going 70+ on the highway. But if it were a rubber tire, you would want to slow down. Using Tesla’s machine learning they are able to tell not just that there is an object, but they are able to tell what kind of object it is. And this is perhaps the most important part, not just seeing an obstacle, but understanding what kind of obstacle it is. This allows the car to continue to learn and develop as the car drives more. It really continues to improve the more driving it does.

My goal this year is to create 50 videos. Each week I'll be talking about a topic related to electric cars, and I would love to hear what you want to learn about. Leave a comment below!
__
#tesla #autonomy #lidar

Want to buy stock in Tesla? Use Robinhood, the free stock trading app. When you use my link to sign up, we'll both get a free stock! http://share.robinhood.com/alexs716

Last video:    • "EV Charging takes FOREVER" | EV Myth...  

Tesla's full event stream:    • Tesla Autonomy Day  

Music from Punch Deck:
   • Punch Deck - Signal in the Noise  

--SOCIAL MEDIA--

Snapchat: asibs21
Twitter: twitter.com/AlexSibila
Instagram: instagram.com/alexsibila/

show more

Share/Embed