There’s a lot of hype surrounding self-driving/autonomous vehicles. Automakers and technology companies are working toward offering various levels of self-driving cars with the goal of being fully autonomous vehicles that require no human intervention. However, for as ambitious as some are about that future just around the corner, the introductory lecture of the Massachusetts Institute of Technology Self-Driving Cars series by leading researcher Lex Fridman, explains just how difficult that will be to achieve.
The introductory lecture, available on YouTube with a host of helpful links in the description, is nearly an hour long and covers a range of self-driving topics. The purpose of the lecture is to look at the autonomous industry as it is today and how it’ll advance throughout the coming year. It looks at several companies dabbling in AV R&D such as Tesla, Ford, Waymo, and Cruise Automation, the General Motors AV subsidiary.
One of the more interesting aspects of the lecture is Fridman’s breakdown of the various sensor technology companies are using in their autonomous vehicles. He breaks down the different types of sensors—Lidar, Ultrasonic, Radar, and Passive Visual—and measures various attributes on a spider chart. The chart shows sensor cost, sensor size, range, resolution, proximity detection, color/contrast, whether it works in rain, snow, fog, and more. It’s a great visual to understand the current state of sensors. However, there are still gaps in their effectiveness.
The lecture does get into the weeds a bit, discussing different types of sensors along with the idea of human-centered autonomous vehicles. The talk also looks at how machine learning and artificial intelligence can assist in developing better autonomous vehicles, but only if there’s a large enough data set.
Autonomous vehicles are coming, but they likely won’t be widely available for some time. Industry experts, after years of ambitious goals and hype, are beginning to realize truly self-driving vehicles will take some time to perfect as many people are still leery of the technology, according to a recent study.
Comments
Instead of seeking the next photo op, perhaps Machete Mary should have attended this conference. She may have learned something for a change.
During the 2020s AEV cars will become popular but who will buy them? Not so many I suspect. AEV cars will be pricey because they need expensive bits especially batteries. And really, the average person, if they can afford an AEV car, will want it to have level 5 autonomy since level 4 leaves much to be desired. With true level 5 autonomy years away, buying an AEV car for oneself will be way in the future, if ever.
Instead of asking “Who will buy an AEV car?” the question to ask is “Who will own an AEV car?”. Expect most of the first AEV cars be owned by companies such as GM and Waymo. They will own them in fleets and use them as taxis. The potential World market for AEV taxis is at least 100 million and whoever owns these first 100 million will potentially have annual net profits above $100 billion for decades. Note that these taxis only have to have level 4 autonomy.
Today I took the wife’s Mini for a nice spring drive, it was wonderful. I drove it, that’s why it was wonderful.