3 ms·

As much as I like LIDAR sensors, and as much as I'd like to see lower cost units, I have a suspicion that they are ultimately going to be unneeded for self-driv

by cr0sh 10y ago

As much as I like LIDAR sensors, and as much as I'd like to see lower cost units, I have a suspicion that they are ultimately going to be unneeded for self-driving vehicles. There's been plenty of demonstrations by various players that self-driving vehicles can get a great amount if not nearly all needed information from camera-based systems, and drive well using them. Other sensors will be needed, of course - and maybe LIDAR can even play a part there; the more data you have, from more sources, the better - this has always been true in robotics. I just think that LIDAR won't play a primary role, but rather as a backup or accessory system component.

05 10y ago
It's a question of reliability. When you market an 'autopilot' that periodically crashes into stationary vehicles, you are probably fine using just a camera. If you, on the other hand, are Google, and want to provide an actual hands-off Level 4 self-driving car, well.. you need an actual 3D point cloud and not just some neural network's educated guess about what the distance is.
- ajross 10y ago
  > not just some neural network's educated guess about what the distance is But that's precisely the mechanism that's been working and accepted since the invention of the directed machine... (Also, just to clarify: a stereoscopic distance determination isn't meaningfully a "guess". That's been a solved problem for a long time.) The goal can't ever be perfect safety. The machines themselves aren't perfectly safe even if they aren't driven into anything. It just has to be measurably safer than a human driver.
  - petra 10y ago
    It's easier to certify a lidar depth map and prove it's reliability, than a complex algorithm that builds depth maps from camera, under all lighting conditions . And in any case , sensor fusion seems to be a common approach, so we'll probably need both lida and cameras.
  - dmd 10y ago
    Well, of course the goal is perfect safety. Whether it's reasonable or desirable to reach that goal is another matter - it's a tradeoff of resources.
- baq 10y ago
  > not just some neural network's educated guess about what the distance is. so, ban humans from driving?
gene-h 10y ago
Now what self driving cars really need are sensors for handling adverse weather conditions. IE sensors for seeing through fog, seeing through rain and snow, detecting ice on the road[0]. Or at the very least a sensor for detecting that things are so bad that the car should just stop. [0]https://www.photonics.com/Article.aspx?AID=6204 https://www.photonics.com/Article.aspx?AID=6204

by cr0sh 10y ago

05 10y ago

It's a question of reliability. When you market an 'autopilot' that periodically crashes into stationary vehicles, you are probably fine using just a camera. If you, on the other hand, are Google, and want to provide an actual hands-off Level 4 self-driving car, well.. you need an actual 3D point cloud and not just some neural network's educated guess about what the distance is.

ajross 10y ago
> not just some neural network's educated guess about what the distance is But that's precisely the mechanism that's been working and accepted since the invention of the directed machine... (Also, just to clarify: a stereoscopic distance determination isn't meaningfully a "guess". That's been a solved problem for a long time.) The goal can't ever be perfect safety. The machines themselves aren't perfectly safe even if they aren't driven into anything. It just has to be measurably safer than a human driver.
- petra 10y ago
  It's easier to certify a lidar depth map and prove it's reliability, than a complex algorithm that builds depth maps from camera, under all lighting conditions . And in any case , sensor fusion seems to be a common approach, so we'll probably need both lida and cameras.
- dmd 10y ago
  Well, of course the goal is perfect safety. Whether it's reasonable or desirable to reach that goal is another matter - it's a tradeoff of resources.
baq 10y ago
> not just some neural network's educated guess about what the distance is. so, ban humans from driving?

gene-h 10y ago

Now what self driving cars really need are sensors for handling adverse weather conditions. IE sensors for seeing through fog, seeing through rain and snow, detecting ice on the road[0]. Or at the very least a sensor for detecting that things are so bad that the car should just stop. [0]https://www.photonics.com/Article.aspx?AID=6204 https://www.photonics.com/Article.aspx?AID=6204