Photo: General Motors
Tesla's and its CEO's stance on LIDAR has become notorious: it would seem standing against the industry's flow is becoming a bit of a theme for Elon Musk as he is the only one to think Level 5 autonomy can be reached without the use of the very accurate laser sensor.
He says the LIDAR has two major drawbacks, with the first being its price. Still, ever since self-driving technology caught speed, the rates have been going down constantly with chances of reaching perfectly affordable levels in just a few years.
The second one is the sensor's inability to see in bad weather, with heavy rain and snow turning it into one big and expensive (see the paragraph above) useless blob on top of the car. The LIDAR has excellent resolution and can map the surroundings with millimetric precision, but only when the atmospheric conditions allow it.
Even so, all the other companies involved in the development of autonomous technology - and there's a lot of them - are betting on LIDAR. In fact, Henrik Fisker tweeted last week that he was installing five of these sensors on his upcoming EMotion electric sedan, saying that was the "minimum number for Level 5 autonomy." While the amount may vary from one company to the other, they all seem to agree on its necessity.
Except for Tesla. Speaking to the Australian media, Scott Miller, GM's director of autonomous vehicle integration, used pretty harsh words to describe Musk's stance. He said the South African entrepreneur was "full of crap" about his claims to reach full self-driving capabilities using only cameras and radars.
“To think you can see everything you need for a level five autonomous [car] with cameras and radar, I don’t know how you do that,” he said. “The level of technology and knowing what it takes to do the mission, to say you can be a full level five with just cameras and radars is not physically possible. I think you need the right sensors and right computing package to do it. [..] you should have redundancy."
If you think about it, it's not sensors that Mr. Miller isn't trusting, but the AI interpreting the information they gather. Humans have been driving using just their eyes for over a century, and despite the many crashes, we seem to be doing a pretty good job. The cameras on a Tesla provide a 360-degree view, meaning nothing is escaping its sight.
The question is whether the company can build a smart-enough AI to interpret this enormous amount of information the right way and in due time. Processing power shouldn't be a problem, but teaching the software how to read every possible situation might be.
With billions of miles of data already under its belt, Tesla is certainly sitting on a small treasure, but its probably jealous competition still doesn't believe it can do it. “Could you do it with less and be less robust? Probably. But could you do it with what’s in a current Tesla Model S? I don’t think so,” Scott Miller concludes.
The second one is the sensor's inability to see in bad weather, with heavy rain and snow turning it into one big and expensive (see the paragraph above) useless blob on top of the car. The LIDAR has excellent resolution and can map the surroundings with millimetric precision, but only when the atmospheric conditions allow it.
Even so, all the other companies involved in the development of autonomous technology - and there's a lot of them - are betting on LIDAR. In fact, Henrik Fisker tweeted last week that he was installing five of these sensors on his upcoming EMotion electric sedan, saying that was the "minimum number for Level 5 autonomy." While the amount may vary from one company to the other, they all seem to agree on its necessity.
Except for Tesla. Speaking to the Australian media, Scott Miller, GM's director of autonomous vehicle integration, used pretty harsh words to describe Musk's stance. He said the South African entrepreneur was "full of crap" about his claims to reach full self-driving capabilities using only cameras and radars.
“To think you can see everything you need for a level five autonomous [car] with cameras and radar, I don’t know how you do that,” he said. “The level of technology and knowing what it takes to do the mission, to say you can be a full level five with just cameras and radars is not physically possible. I think you need the right sensors and right computing package to do it. [..] you should have redundancy."
If you think about it, it's not sensors that Mr. Miller isn't trusting, but the AI interpreting the information they gather. Humans have been driving using just their eyes for over a century, and despite the many crashes, we seem to be doing a pretty good job. The cameras on a Tesla provide a 360-degree view, meaning nothing is escaping its sight.
The question is whether the company can build a smart-enough AI to interpret this enormous amount of information the right way and in due time. Processing power shouldn't be a problem, but teaching the software how to read every possible situation might be.
With billions of miles of data already under its belt, Tesla is certainly sitting on a small treasure, but its probably jealous competition still doesn't believe it can do it. “Could you do it with less and be less robust? Probably. But could you do it with what’s in a current Tesla Model S? I don’t think so,” Scott Miller concludes.
Google News