The Car Guide is pleased to welcome techno columnist Mathieu Roy among its collaborators. Mathieu will approach techno from all angles in a series of new chronicles. For this first text, we address the issue of the famous autonomous vehicles.
They’re called self-driving cars, but let’s face it, most really aren’t.
To be more accurate, we must change the vocabulary: we are talking about driving assistance. As long as the human will take over part of the work, it is difficult to speak of autonomous driving.
Car manufacturers are not the most innovative in this area either. We are not up to date with companies that work in techno. Moreover, manufacturers often make alliances with technology companies to take advantage of their expertise.
Obviously, we think of Google and its WAYMO division, which has been working for years to implement the autonomous car, in all kinds of conditions and obviously in large urban centers. There have also been more than 4.3 million kilometers traveled in autonomous mode in San Francisco alone. There, we even implemented a principle of taxi without driver.
A funny experience
On a more personal note, I had been thrown on my ass when I got into a Yandex car in Las Vegas. A whole experience, which I had also shared on Facebook:
They exist
Tesla is also working very hard to automate its cars. GM is an investor in Cruise, which will also have a taxi fleet in San Francisco. So self-driving cars do exist.
They exist, these cars, and they see much better than us because we plant a lot of sensors in them. It includes cameras, GPS systems, radars, lidars, a software solution to manage artificial intelligence and computer algorithms to merge data from all these sensors. Moreover, in Quebec, the company LeddarTech, which specializes in this type of major analysis.
Cars learn and they learn fast. One of the ways the self-driving car learns is through the sharing of information. The cars “learn” in a swarm by communicating to the others each small error in the course so that it is not repeated. Learning and improving a car can thus benefit the entire fleet. This is why progress is rapid.
Photo: Waymo
The five levels
There are five levels of automation.
Level 1: Driving assistance. We are talking here about technologies as simple as adaptive cruise control. A large portion of new vehicles sold in Canada offer this type of technology today.
Level 2: Partial automation. Electronic systems control both the steering and acceleration functions, but the person behind the wheel remains essential to the proper functioning of the vehicle.
Level 3: Conditional automation. When the vehicle is in motion, it can technically drive itself, but it can ask the person behind the wheel to take over control at any time.
Level 4: High automation. The vehicle can be operated without human intervention, but it remains equipped with pedals and a steering wheel. The human can regain control if desired or if an emergency maneuver is required.
Level 5: Full automation. The vehicle operates by itself and no human intervention is required.
Often technology moves faster than our ability to accept these same technologies.
In the meantime, the autonomous car continues to be deployed… according to the latest news, the company Argo AI is going to run its cars without a driver in Miami and Austin in the United States. Here is a video example: