How do driverless cars work?
Prompt: Sideview of a modern sleek driverless sedan on a city street
Imagine you're playing a video game with a car that drives itself - driverless cars are like those, but for real! They have special "eyes" called sensors that see everything around them like roads, cars, and even people. But what makes them truly amazing is their brain. This brain is powered by artificial intelligence (AI) which uses all the information from the sensors to make decisions in a blink: steering, braking, even following traffic lights, just like a real driver. Driverless cars are still learning, so we need to remember to keep an eye on things until they're truly masters of the road!
Humans sense the world through five basic senses: sight, hearing, smell, taste, and touch. Using our senses, we can understand objects, scenarios and terrain.
As we move around in spaces we can recognise a door, a step, an incoming vehicle, someone walking in front or behind us, water that has been left on the floor/ground, etc. We can also make decisions and act with our body to slow down/speed up, turn left or right, stop/start, and so on.
Autonomous or driverless cars also have to do exactly the same steps to move around in space safely. They need to sense their surroundings and so are fitted with different types of “sensors” which are electro-mechanical devices that can collect images like a camera, sense heat like a home thermometer, sense sound like a microphone, etc.
Driverless cars have a number of cameras which continuously take video images which are then processed by a computer within the car. The program in the computer uses image recognition to understand whether it detects a tree, street light, a signal (red/green/yellow) or a person crossing the street, etc. It can also detect speed limits marked on signs posted on the side of the street as well as street names. The computer can also calculate the distance between itself and these objects.
Once the computer has a good “picture” of what is in the surrounding environment of the car and the direction in which it needs to go, supported by a positioning navigator, it uses a set of rules on what to do and what not to do for that scenario. For example, one rule can be “stop the car when it is too close to an object in front”. But how does it know if it is too close? It calculates its action based on the speed it is traveling and the distance to the object.
There can be hundreds of such rules - driving and traffic rules - such as stop when the signal is red, stay within speed limits, stay within the line, etc. Artificial intelligence helps driverless cars remember and act on these rules.
