Chapter 1: What Is a Self-Driving Car, Anyway?

Imagine you’re sitting in the backseat of your family’s car on a road trip. Your dad is in the driver’s seat, one hand on the steering wheel, eyes fixed on the road, occasionally tapping the brakes or turning the wheel. The car moves because your dad decides when to speed up, slow down, or turn. But what if there was no one in the driver’s seat at all? What if the car could make all those decisions all by itself, using sensors and computers instead of a human brain? That’s exactly what a self-driving car is!

[插图1：一家人坐在车里，爸爸在开车，旁边有个小气泡，里面画了一辆没有司机的车，在路上自己开，标注“这就是自动驾驶车的魔法：不用人开，自己就能跑！”]

What Makes a Car “Self-Driving”?

To put it simply, a self-driving car is a regular car that has extra special parts added to it, so it can drive without a human telling it what to do all the time. Regular cars only have one “brain” (the human driver) and one set of “eyes and ears” (the driver’s eyes, ears, and the car’s basic mirrors). Self-driving cars have a super powerful computer as their brain, plus lots of extra sensors that act like super-powered eyes, ears, and even a sense of touch! These sensors can see everything around the car 360 degrees, day or night, rain or shine, and send all that information to the computer super fast. The computer then decides when to speed up, slow down, turn, or stop, just like a human driver would.

> [!DID YOU KNOW?] > Self-driving cars are also called “autonomous vehicles” or “AVs” for short! “Autonomous” is a fancy word that means “able to work on its own without help from humans.”

You might have already seen a self-driving car without even noticing! They look almost exactly like regular cars, except they usually have a funny-looking box on the roof (that’s the laser radar, we’ll learn about that in Chapter 2!) or special stickers on the side. If you look closely next time you’re on the sidewalk, you might spot one!

How Are Self-Driving Cars Different from Regular Cars?

Let’s compare a regular family car and a self-driving car side by side to see what makes them special:

| 特点 | 普通汽油/电动车 | 自动驾驶车 | |------|----------------|------------| | “大脑” | 人类司机 | 车载超级电脑 | | “感官” | 司机的眼睛、耳朵，车的基本后视镜 | 摄像头、激光雷达、毫米波雷达、超声波雷达、GPS，360度无死角 | | 反应速度 | 看到障碍物到踩刹车需要0.5-1秒 | 0.1秒以内，比人类快10倍 | | 会分心/疲劳吗？ | 会，看手机、聊天、开久了会累 | 不会，24小时一直盯着路，不会累 | | 需要驾照吗？ | 必须要有驾照的人开 | L4/L5级别的完全自动驾驶车，乘客不需要驾照 | | 会酒驾/超速吗？ | 可能会 | 永远不会，程序里设定了遵守交通规则 |

[插图2：普通车和自动驾驶车的对比图，左边是普通车，爸爸在开车，旁边标注“1个大脑，2只眼睛”；右边是自动驾驶车，车顶有个盒子，里面有个电脑，周围画了好多小眼睛，标注“1个超级大脑，360度眼睛！”]

The biggest difference is that self-driving cars never get tired, never get distracted by their phone, never drive too fast, and never drive after drinking. Remember, 94% of all car crashes are caused by human mistakes, like texting while driving, running a red light, or driving when tired. If self-driving cars work perfectly, they could make roads way safer for everyone!

> [!DID YOU KNOW?] > Waymo, one of the first companies to make self-driving cars, has driven more than 20 million miles (32 million kilometers) on public roads as of 2024. That’s like driving around the entire Earth more than 800 times!

A Quick Trip Back in Time: The History of Self-Driving Cars

People have been dreaming of cars that drive themselves for almost 100 years! Let’s go on a time trip to see how we got from silly sci-fi ideas to the self-driving cars we see today.

1920s-1950s: Sci-Fi Dreams and First Experiments

The first idea of a self-driving car showed up in a 1926 German sci-fi movie called *Metropolis*. The movie showed futuristic cars driving around a huge city all by themselves, no drivers needed! Back then, cars didn’t even have seatbelts or airbags, so a self-driving car was just a silly fantasy, right?

[插图3：1926年电影《大都会》的自动驾驶车场景，复古科幻风格，标注“90年前，电影里的人就已经想象出不用人开的车啦！”]

But in the 1950s, a company called RCA decided to try to make that fantasy real. They built a small car that could drive all by itself—but only if there were special wires buried under the road! The car followed the electric signals from the wires, just like a toy remote-control car. It worked perfectly, but you couldn’t drive it on normal roads, because normal roads didn’t have those wires under them. It was a cool experiment, but not very useful for real life.

1980s: The First Car That Could Drive on a Real Road

It wasn’t until the 1980s that someone finally built a car that could drive on a normal, public road without any special wires. A team of engineers at Stanford University built a funny-looking little vehicle called the Stanford Cart. It looked more like a lawnmower with a big computer stuck on top than a real car! But in 1989, it did something no one had ever done before: it drove 600 feet (about 180 meters) all by itself on a regular road next to the Stanford campus. There was a safety driver sitting in the seat next to it, just in case, but the car was controlling the steering and the brakes all on its own. It was very slow—only about 1 mile per hour, so you could run faster than it!—but it was the first real self-driving car that could work on normal roads.

[插图4：1980年代的Stanford Cart，黑白照片，看起来像加了电脑的割草机，旁边标注“这台开得比爷爷骑自行车还慢的机器，是第一个能在普通公路上自己开的自动驾驶车哦！”]

At the same time, a team at Carnegie Mellon University was working on their own self-driving van called NavLab. They added a big computer, a camera, and special sensors to a regular minivan, and it could drive itself on the highway, recognizing road signs and staying in its lane. It was way faster than the Stanford Cart, but it still needed a human driver to help sometimes.

2000s: The DARPA Grand Challenge

In the 2000s, the U.S. Department of Defense got interested in self-driving cars. They thought if cars could drive themselves, they could send supplies to soldiers in dangerous war zones without putting human drivers at risk. So they held a huge competition called the DARPA Grand Challenge, with a $1 million prize for the first team that could build a car that could drive 142 miles (about 230 kilometers) through the Mojave Desert all by itself.

The first competition in 2004 was a total mess! 15 teams entered, but none of

could even make it 10 miles (16 kilometers) before crashing, getting stuck in the sand, or breaking down completely. The best-performing car only drove 7 miles (11 kilometers) before it ran into a big rock and stopped. The crowd of engineers and spectators laughed and cheered, but everyone knew they had learned a lot from the failure—and they were already planning to try again next year.

<figure> <插图5：[彩色插图] 2004年DARPA大挑战的沙漠赛道，几辆造型奇特的自动驾驶车有的停在沙地里，有的撞在岩石上，旁边站着捂脸的工程师，气泡里写着“2004年的比赛：没有一辆车跑完全程，全是‘翻车现场’！”> <figcaption>2004年DARPA大挑战的“翻车现场”</figcaption> </figure>

The second DARPA Grand Challenge in 2005 was a total game-changer. This time, 23 teams entered, and for the first time ever, multiple cars finished the full 142-mile (230-kilometer) desert course. The winning car was named Stanley, built by a team of students and engineers from Stanford University. Stanley looked like a regular SUV with a big box of sensors and computers stuck to its roof, and it drove the whole course in just 6 hours and 52 minutes—20 minutes faster than the second-place car, built by Carnegie Mellon University. What made Stanley so special? It didn’t panic when it ran into small cacti, dirt mounds, or even a fallen tire on the road. It used its sensors to spot the obstacle, calculate a safe path around it, and keep driving without stopping. After the race, Stanley was put on display at the Smithsonian National Air and Space Museum in Washington, D.C. If you ever visit the U.S. capital, you can go see this famous self-driving car in person!

<figure> <插图6：[彩色插图] 斯坦利SUV冲过DARPA 2005年终点线的瞬间，旁边围着欢呼的工程师团队，车顶的传感器盒子亮着蓝灯，标注“冠军诞生！