To understand self-driving cars, we first need to define them. Fully autonomous cars are cars that can take in information about their environment and use it to transport people or objects from point A to point B without human intervention. This type of autonomous car is known as a level 5 autonomous car. There are 6 levels of self-driving cars:
As of right now, there are no level 5 self-driving cars. In fact, even Tesla cars are only classified as level 2 autonomous cars, meaning the car only assists the driver. The driver still does most of the driving.
So how do self-driving cars work?
One common type of ML model used to create self-driving cars is Reinforcement Learning. Reinforcement Learning programs take in information about an environment and use that information to take an action to get closer to its goal. The program then gets feedback on how successful their action was and gets more information about the environment to take another action.This process keeps repeating until the program reaches its goal. Here is a diagram of how it works:
Here is a video explaining Reinforcement Learning:
Another way to create the program for self-driving cars is to use Supervised Learning.