Since autonomous vehicles do not have a human driver, how do they navigate the world around them? The answer is they rely on a variety of different sensors to mimic human capabilities. There are four main types of sensors in use today: camera, radar, lidar and sonar.
Cameras are by far the most pervasive sensor and you cannot find an autonomous vehicle that doesn’t have many (often 8 to 12) of them pointed in different directions. A very simple function, among many, that a camera serves is to read or interpret road signs and traffic signals. Radar is the second most prevalent sensor, mainly because it functions well in poor visibility and the technology is relatively inexpensive. Radar is mostly used to determine the distance of objects (e.g., another vehicle) from the car. Sonar is a cousin of radar (uses sound waves instead of electromagnetic waves) but appears to be losing favor among vehicle manufacturers mainly because radar can do most of what sonar can at a lower price point.
The final sensor we will discuss is lidar which stands for “light detection and ranging”. Lidar uses laser light and its reflection to provide a 360-degree, 3D map of the surroundings of the vehicle. If you have seen an autonomous vehicle, these are the spinning modules one often sees on the roof or side panel. While lidar is tremendously helpful to map the complete environment around the car, it is by far the most expensive sensor and there is a very strong push by technologists to reduce the cost drastically.