It’s already 2025, why aren’t self-driving cars close to every household? !

Welcome to Science Popularization China’s special winter vacation column “ High-tech Classes for Children ”!

As one of the most cutting-edge technologies today, artificial intelligence is changing our lives at an astonishing speed. From smart voice assistants to driverless cars, from AI painting to machine learning, it has opened up a future full of infinite possibilities for us. This column will use videos and text to explain to children the principles, applications, and profound impact of artificial intelligence on society in an easy-to-understand way.

Come and start this AI journey with us!

Since 2015, there has been a lot of news about self-driving cars. Technology giants and auto giants have entered the market and want to develop their own self-driving cars. At that time, many people believed that in a few years, the streets would be full of self-driving cars. Humans would not even need to take a driver's license test.

But it’s been almost 10 years, why haven’t self-driving cars arrived yet?

In this episode, we will take a closer look at the past and present of autonomous driving .

Development of autonomous vehicles

People have been working on self-driving cars for longer than you might think.

The earliest self-driving car appeared in 1925 , but there were no computers at that time, so how could it be self-driving? In fact, the self-driving cars at that time did not rely on AI technology, but on wireless remote control technology. Someone was controlling the car remotely, and there was no need for the person in the car to drive, just like we play with remote control cars.

In fact, as long as the person driving the car is replaced by a computer, automatic driving can be achieved. This is easy to say, but it is too difficult to achieve. In the 1950s and 1960s, people thought of another solution, that is, let the road control the car .

For example, in 1957, RCA in the United States tested an autonomous car on a highway. The highway had been modified, with many sensors that could detect the position and speed of the car and send signals to it for remote control. At this time, the people responsible for observation and control were no longer humans, nor the car, but various sensors on the road.

The road surface has also been modified, and the cars will drive along the set routes, which is a bit like a tram. When computer vision and AI technology were not yet mature, this was indeed a very creative idea.

In the 1960s and 1970s, machine vision was developing rapidly, making it possible for cars to observe their surroundings. For example, Stanford University developed a Stanford car that can analyze photos taken by cameras in real time and identify roads, which is more in line with our understanding of autonomous driving.

In the 1980s, modern self-driving cars appeared. In addition to video cameras to identify the road, self-driving cars at this time also added lidar to help the car see the surrounding situation clearly. With the help of these technologies, self-driving cars can drive on roads without pedestrians and other vehicles.

Most notably, in the 1980s, Carnegie Mellon University built a NAV-1 self-driving car. This car has begun to use artificial neural networks and back propagation algorithms to train the car to recognize roads. The basic principle is very similar to modern self-driving cars. In the 1990s, with the development of neural networks, self-driving cars have been able to challenge the roads in daily life.

For example, in 1994, a self-driving car developed by Mercedes-Benz traveled more than 1,000 kilometers on the congested highway in Paris. Although human intervention was still required when the road conditions were particularly complicated, the self-driving car was able to automatically overtake and follow the car.

In 1995, Carnegie Mellon's NAV-5 completed a 5,000-kilometer challenge. During this process, the car's steering wheel was controlled by AI for 98% of the journey, but for safety reasons, the accelerator and brakes were still operated by humans.

In 1996, Italian scientists also achieved 1,900 kilometers of AI autonomous driving. During this process, 94% of the roads were driven by AI autonomous driving.

The Hardest 1%

After 2000, the combination of convolutional neural networks and GPUs has made computer vision advance by leaps and bounds. Coupled with increasingly accurate sensors, autonomous driving technology has also exploded. Google's Waymo cars and Tesla's Autopilot system have been launched one after another. In 2017, Baidu's Apollo driverless car also flew on Beijing's Fifth Ring Road.

All this makes people feel that self-driving cars will soon enter thousands of households. Why don’t we hear any news about self-driving now? Has self-driving been cancelled?

No.

Copyright images in the gallery. Reprinting and using them may lead to copyright disputes.

Although self-driving cars were able to handle more than 90% of situations on their own before 2000, the last 10%, or even 1%, are the most difficult .

For example, even if the computer can accurately identify road conditions at 99.99%, the remaining 0.01% may still cause major problems. For traffic safety, no mistakes are allowed. In addition, most countries do not have comprehensive legislation on autonomous driving. These are issues that need to be resolved before autonomous vehicles are put on the road.

So has the development of self-driving cars stopped? No, in closed areas such as airports and mines, self-driving cars can still play an important role. For example, in our country, many airports have adopted low-speed driverless cars. For example, unmanned sweepers, unmanned tractors, unmanned patrol cars, etc. In dangerous and closed areas such as mines, driverless cars can also be of great use. In addition, in some relatively closed parks, there are still low-speed driverless cars being tested and operated.

Autonomous driving has been a dream of people for more than 100 years. I believe that one day the roads will be full of autonomous vehicles. But before that day comes, we must be fully prepared.

Planning and production

This article is a work of the Science Popularization China-Creation Cultivation Program

Produced by: Science Popularization Department of China Association for Science and Technology

Producer｜China Science and Technology Press Co., Ltd., Beijing Zhongke Xinghe Culture Media Co., Ltd.

Author: Beijing Yunyuji Culture Communication Co., Ltd.

Reviewer: Qin Zengchang, Associate Professor, School of Automation Science and Electrical Engineering, Beihang University

Planning丨Fu Sijia

Editor: Fu Sijia