Self-driving car definitions and technology

Self-driving car overview and terminology

A self-driving car (also called an autonomous, driverless, robotic, or robo-car) is a vehicle designed to operate with reduced or no human input. The term “self-driving” does not have a single universally agreed definition and is often influenced by vendor branding. In practice, systems are typically built from sensor suites (commonly LiDAR and cameras), software (including AI and neural networks), and supporting technologies such as GPS and established ADAS engineering, but they still face limitations that can lead to accidents. Because capabilities vary by operating conditions, terminology often centers on the operational design domain (ODD)—the specific environments and scenarios where the system is intended to work safely. Key distinctions include advanced driver-assistance systems (ADAS), which automate limited driving functions while requiring human supervision, versus automated driving systems (ADS) that can take on more driving tasks. “Autonomy” and “automation” are also differentiated: autonomy implies the system is under vehicle control rather than the driver, while automation is more function-specific and may leave broader decisions to the human. To standardize communication, multiple classification and terminology schemes are used. SAE International’s J3016 defines automation levels (0 through 5) based largely on the driver’s role and responsibility, with higher levels shifting more safety-critical responsibility to the system. Mobileye uses a “hands/eyes” and “no driver” vocabulary (e.g., eyes-on/hands-off, eyes-off/hands-off, no driver) to describe human monitoring requirements. British and European definitions further shape how terms like “automated,” “autonomous,” and “driving itself” are legally interpreted, and they emphasize that safe operation depends on meeting defined conditions and requirements.

0/2 modes complete

Operational design domain (ODD) limitations

0/2 modes complete