Put simply, LiDAR is just radar using a laser light instead of radio waves to scan what is in front of it. LiDAR stands for Light Detection and Ranging and was the next big iteration in remote sensing and imaging thirty years ago. Today it is everything from airplanes, to cars, to ground survey systems.
Here is the 9 minutes video version of this article:
Not to leave anyone out of the party, sonar bounces sound waves off of objects to figure out what is near it.
Most mid-range and luxury vehicles today have both radar and sonar sensors. Radar is cheap, reliable and most frequently hidden in the front grill to tell the cars computer systems what is in front of it, where as sonar are those tiny round wounds in your bumpers are used to tell you how many more inches you have before you hit something when parking.
Radar’s biggest technical advantage over LiDAR is that radar can see through fog, dust, normal snow and rain. Light waves just don’t pentrate those blockages well. Even better, radar is cheap, effective and has few moving parts to fail.
So, what can LiDAR do that radar can’t?
In a word, both. If you have seen a truly autonomous car you will have seen that funny looking bucket spinning furiously on top of the card is LiDAR. If you think that is ugly and expensive, you are right. LiDAR systems for cars today are mostly bespoke and cost between $5,000 and $10,000 per unit. That kind of cost, complexity and visual clutter just isn’t going to ever be part of a mass market vehicle… or is it? Has something changed?
Take a quick look at all of these pictures and think about what is common to all of them but not Tesla:
Clearly Elon Musk is the Howard Hughes of this century. A brilliant, insightful, free thinking individual who is not only capable of thinking outside of the box, but demands the same from his staff. He is nobodys fool and when he makes forceful statements about technology and business you had think long and hard before challenging him. Here is what Musk said in April 2019:
“… LiDAR is a fools arrange and anyone relying on LiDAR is doomed. (They are) expensive sensors that are unnecessary.”
Musk beleives that cameras and radar information fed into artificial intelligence system can provide the same information at a tiny fraction of the cost of LiDAR. He is no doubt correct. However there are two massive issues he has not addressed:
Aptiv (formerly Delphi a GM electronics company), LedderTech, and others have drastically changed LiDAR. LiDAR is now tiny, solid state and inexpensive. LeddarTech now enables LiDAR is vehicles for $500 and within two years that will drop to just $300 per car.
Musk wrongly frames the argument as radar vs lidar. The solution to autonomous driving is radar + lidar + cameras + big data + vehicle to vehicle communication + artificial intelligence.
For autonomous vehicles to succeed, they must be at LEAST twice as safe as human drivers. If not, every exceptional autonomous mistake will be held up as an example of why they need to be either banned outright in some jurisdictions or needlessly delayed for years. That is not good for anyone.
LiDAR and Radar are certainly competitors but in the case of autonomous vehicles they will work together to keep us safe.
Do you really want to drive down the road in a glass and steel box that does not have a serious back-up system?
I have a Tesla Model 3 now while I truly love autopilot (actually auto steer) I can tell you from first hand experience that the number of self driving mistakes it makes is terrifying. Tesla autosteer is acceptable on most (certainly not all) highways as long as the driver is casually paying attention but it can’t figure out even the most basic navigation in the city. Tesla’s don’t avoid pot holes, road debris or even gravel; they can’t figure out what to do when a lane ends which FREQUENTLY results is the vehicle lurching dangerously in to or out of a lane; and they certainly can’t navigate novel road situations like construction barriers, human flaggers, police / ambulances / snow plows, or accident scenes.
Thus far autonomous vehicle manufactures have been given largely free reign to develop and test their wonderful life saving machines on public roads. This will change. Currently there is a serious distrust among both politicians and citizens of both big data and big tech companies. Tesla likes to promote itself, especially to politicians and the stock market, as a tech company. That has worked in their favor for years but we will not be at all surprised to see a few dozen autonomous vehicle mistakes result in high profile deaths, leading a regulatory response that dictates what technologies MUST be used in autonomous vehicles.
That will not be a good day for innovation, tech companies, or car companies but it just might be what consumers need to believe autonomous vehicles are safe.
This website uses cookies.
View Comments