Learn

Under the hood

Plain-language explainers on autonomous vehicles: how the technology works, where it came from, and the debates around it. Pick a subject, or read them all.

New to the jargon? Start with the glossary →

Foundations2014SAE LevelsThe SAE Levels are a six-tier scale, from Level 0 to Level 5, defined in SAE standard J3016, that classifies how much of the driving a vehicle can do on its own. Levels 0 to 2 assist a human driver who stays responsible; Levels 3 to 5 let the vehicle drive itself under defined conditions.Read →
History & approaches2015The Google approach: all or nothingGoogle's self-driving car project, now Waymo, pioneered a strategy of skipping the partial, human-supervised steps of the SAE ladder. After early tests showed that people cannot reliably supervise a car that mostly drives itself, the team committed to a fully automated driver with no human fallback. The stance is often summed up as all or nothing.Read →
History & approaches2004The DARPA Grand ChallengeThe DARPA Grand Challenge was a series of US government prize races, held in 2004, 2005, and 2007, that asked robotic vehicles to drive themselves with no human aboard. Widely seen as the big bang of the modern self-driving car, it produced the people and ideas behind nearly every major program that followed.Read →
How it works2016End-to-end self drivingEnd-to-end self driving uses a single neural network that takes raw sensor data and outputs driving commands, the steering, accelerator and brake, learning to drive directly from examples. It contrasts with the modular approach, which breaks driving into separate perception, prediction, planning and control stages, many of them governed by hand-written rules.Read →
How it works2008Sensing & PerceptionSensing is how a self-driving car perceives its surroundings. It combines several sensor types, cameras, radar, lidar and ultrasonic sensors, with satellite positioning and prior maps. Each has different strengths and blind spots, so most systems fuse them for redundancy, while a few rely mainly on cameras. Together they answer where the car is and what is around it.Read →
Policy & regulation2011Regulatory complianceRegulatory compliance is the set of laws, permits and safety rules a self-driving car must meet to test or operate on public roads. There is no single global standard. In the United States it is largely a state-by-state patchwork, layered on top of federal vehicle safety rules, and it differs again across Europe and Asia.Read →
History & approaches1977The first self-driving carThere is no single first self-driving car; it depends on what you count. A radio-controlled car drove through New York in 1925, test cars followed wires buried in the road in the 1950s, and the first car that steered itself by watching the road with cameras was built in Tsukuba, Japan, in the late 1970s.Read →
How it works2018RobotaxisA robotaxi is a ride-hail vehicle driven by an automated driving system, with no human driver, that operates only inside a defined area and set of conditions called its operational design domain. Riders book it in an app like any taxi. The software drives, and remote staff can advise it but do not steer it.Read →
History & approaches2026Waymo vs TeslaWaymo and Tesla are making opposite bets on how to reach self-driving. Waymo runs driverless robotaxis that combine lidar, cameras, radar and detailed maps, expanding city by city. Tesla relies on cameras and neural networks alone, across millions of customer cars and a growing robotaxi service, betting that scale and data will let it drive anywhere.Read →
How it works2025Autonomous truckingAutonomous trucking uses automated driving systems to drive heavy freight trucks, starting on highways between terminals near interstates. Since 2025 a few companies, led by Aurora in Texas, have hauled paying freight with no one driving, while people still handle local pickup and delivery at each end of the trip.Read →
Foundations1995ADASADAS, short for advanced driver assistance systems, are features that help a human drive, such as automatic emergency braking, lane keeping and adaptive cruise control. They sit at SAE Levels 0 to 2: however capable they feel, the person behind the wheel is still the driver and must stay attentive at all times.Read →
How it works2026AI foundation modelsA foundation model for driving is a single large neural network, pre-trained on vast amounts of driving video, maps and language, that a company then adapts to many tasks: driving the car, simulating the world around it, and judging its decisions. Since 2024 the leading developers have rebuilt their software around such models, borrowing the recipe of large language models.Read →
FoundationsGlossaryA plain-language glossary of the terms engineers, regulators and reporters use about self-driving cars, from ADAS and ODD to ego vehicle and disengagement. Each entry is a short, self-contained definition, consistent with the SAE J3016 standard where it defines the term, with a link to the lesson that goes deeper.Read →