Self-driving car
Adapted from Wikipedia · Discoverer experience
A self-driving car, also known as an autonomous car, driverless car, or robotic car (robo-car), is a car that can operate with little or no help from a person. These cars use special tools like LiDAR (which measures distances using light) and cameras to "see" the world around them. They also rely on GPS and smart computer programs to make decisions about driving.
As of 2026, no self-driving car can drive perfectly in every situation. Companies like Waymo have started offering driverless taxis in some areas, but these cars still face challenges. They need very detailed maps and lots of testing to stay safe on busy roads. There are also questions about how to make sure these cars are secure and reliable, and how laws should handle them.
People’s feelings about self-driving cars are mixed. Some are excited about the idea of not having to drive, while others are worried about safety. As these cars develop, they could change how we build cities and how we travel, but many challenges still need to be solved.
History
Self-driving cars, also called autonomous or driverless cars, can operate with little or no help from a person. These vehicles use sensors and computers to navigate roads and make decisions. The idea of self-driving cars has been around for many years, with early concepts appearing in the late 1920s. Over time, technology has improved, leading to cars that can drive themselves in many situations today. Some companies now use self-driving cars as robotaxis for ridesharing services. However, there is no single, agreed-upon definition for what counts as a "self-driving" car, and the term is often used in advertising.
Definitions
Organizations like SAE International have tried to create rules for calling a car "self-driving," but there is no one agreed-upon definition. Different companies use different words and features, so it can be confusing.
Some cars can drive themselves only in certain places or conditions. For example, some cars can drive without driver help only on highways, while others need a driver ready to take over at any time. The words used, like "full self-driving" or "autopilot," don’t always mean the same thing. In the United Kingdom, a car is called self-driving if it is officially registered that way, not just because it has certain features. Some experts warn that using words like "autonomous" in ads might make people think the car can drive itself more than it really can.
| Mode | Level | Summary | Description | Responsibility for | |||
|---|---|---|---|---|---|---|---|
| Direction & speed | Monitoring environment | Fallback | |||||
| n/a | 0 | No Automation | Full-time performance by the driver of all aspects of driving, even when "enhanced by warning or intervention systems" | Driver | Driver | Driver | |
| Some | 1 | Driver assistance | Driving mode-specific control by an ADAS of either steering or speed | ADAS uses information about the driving environment; driver is expected to perform all other driving tasks. | Driver and system | ||
| 2 | Partial automation | Driving mode-specific execution by one or more ADAS for both steering and speed | System | ||||
| 3 | Conditional automation | Driving mode-specific control by an ADAS of all aspects of driving | Driver must appropriately respond to a request to intervene. | System | |||
| Many | 4 | High automation | If a driver does not respond appropriately to a request to intervene, the car can stop safely. | System | |||
| All | 5 | Full automation | System controls the vehicle under all conditions and circumstances. | ||||
Technology
Main article: Vehicular automation
Self-driving cars use special technology to "see" and understand their surroundings. They have sensors like cameras and radar that collect information about the road, other cars, and traffic signs. This information helps the car decide how to move safely.
These cars also use maps to find their way from one place to another. Some maps are very simple, while others are very detailed. The car can use these maps together with its sensors to navigate. Artificial intelligence helps the car process all this information and make smart decisions while driving.
Challenges
Self-driving cars, also known as autonomous cars, need very advanced software and maps to work safely in all kinds of weather and road conditions. They must also deal with other cars, obstacles, and confusing traffic rules. This makes them very tricky to develop.
There are also concerns about cost, safety, and how people feel about using them. Some companies have been accused of misleading people by calling their cars "self-driving" when they still need a human driver nearby. There are also worries about security, like hackers trying to steal information from these cars. Additionally, self-driving cars might change jobs, as many people work as drivers, and this could affect many lives.
Testing
Self-driving cars, also called autonomous cars, can be tested in many ways. They might be tested using computer programs, in special areas made just for testing, or on regular streets. When testing on public roads, there are often rules to follow. For example, in New York, a person must sit in the car and be ready to help if needed.
In the 2020s, many new ways to test self-driving cars were developed. Different companies used different rules for when a car needed a person to take over. In 2021, a group called WP.29 suggested a way to test these cars. Later that year, tests were done in Hamburg, Germany, to try out new car technology on normal streets. In 2022, a safety test method was created, and companies like Nissan and Waymo showed how their cars could avoid crashes. Researchers also studied how people react when they see a self-driving car without anyone inside.
Safety
A meta-analysis published in Nature Communications in 2024 looked at safety data for self-driving cars and cars driven by people. The study found that self-driving cars are often safer, especially in bad weather like heavy rain or fog. However, they can have more trouble seeing at dawn and dusk.
Some self-driving cars, like robotaxis, work very well on their own, while others use helper systems to assist drivers. These helper systems can make driving easier but still need a person to watch and steer when needed.
Because self-driving technology is new, there have been some problems. For example, in 2016, a Tesla car using its helper system had an accident in China. In 2018, an accident in Arizona involved a self-driving car from Uber. These events show that more testing and learning are needed to make self-driving cars completely safe for everyone.
Public opinion surveys
In 2011, a survey of people in the United States and the United Kingdom showed that almost half felt comfortable using a car that drives itself. In another 2012 survey, about a third of car owners were interested in buying a car that could drive itself, but that number dropped if the car cost more money.
More recent surveys show that people have mixed feelings about self-driving cars. Some worry about safety or how the cars might share information, while others see benefits. In 2022, only about a quarter of people worldwide said they would feel safe in a self-driving car.
Regulation
Main article: Regulation of self-driving cars
See also: Regulation of algorithms
Self-driving cars, also called autonomous or driverless cars, can operate with little or no help from a person. Right now, there isn’t a clear, agreed-upon definition for what counts as a “self-driving” car, and companies sometimes use the term in their advertising. Rules and laws about these cars are still being developed to make sure they are safe and work well for everyone.
Commercialization
See also: History of self-driving cars § 2020s
Self-driving cars can change levels depending on the task and situation. For example, a car might have features that work at a higher level in some situations but not others.
As of 2023, most cars available for buying have Level 2 self-driving features. Some companies have reached higher levels, but only in certain places. Even cars below Level 5 offer many helpful features.
Level 2 – Partial automation
See also: Lane centering § Sample of level 2 automated cars, and List of self-driving system suppliers § Date of first public road driverless operation
Level 2 features are common in many cars. In the US, half of new cars have help for steering and speed control.
Ford started offering a service called BlueCruise in 2022. It helps with staying in lane, reading street signs, and driving hands-free on many highways. In April 2023, BlueCruise became approved for use in the UK on certain motorways.
Tesla's Autopilot and its Full Self-Driving systems have been in all Tesla cars since 2016. These systems help with driving on highways and streets, changing lanes, and avoiding crashes, but the driver must always be ready to take control.
Level 3 – Conditional automation
As of April 2024, only two car makers sold Level 3 cars: Honda in Japan, and Mercedes in Germany, Nevada and California.
Mercedes Drive Pilot has been available on the EQS and S-class sedan in Germany since 2022, and in California and Nevada since 2023. It can only be used under certain conditions, like on freeways mapped by Mercedes.
As of 2026, the only level 3 cars in China are Deepal SL03 and Arcfox αS.
Level 4 – High automation
In 2024, companies like Waymo started offering self-driving taxi services in some parts of the US. These cars drive completely on their own but do not yet cost less than regular cars to operate.
In April 2023 in Japan, a new rule allowed Level 4 self-driving cars. ZEN drive Pilot Level 4 made by AIST operates there.
Images
Related articles
This article is a child-friendly adaptation of the Wikipedia article on Self-driving car, available under CC BY-SA 4.0.
Images from Wikimedia Commons. Tap any image to view credits and license.
Safekipedia