CPC class G05D is where the patent system files the brain that steers an autonomous machine. In the Cooperative Patent Classification scheme maintained jointly by the USPTO and the European Patent Office, G05D is the subclass titled "SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES." That title is broad, because G05D covers control of many physical variables, but its most important group for robotics and autonomy is the very first one, G05D 1/00, whose official title is "Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots." That single group is the classification home of self-driving cars, autonomous drones, mobile robots, and any system that decides and controls where a vehicle is and where it is heading.

The breadth of G05D 1/00 is deliberate and worth internalizing: it spans land, water, air, and space. A robotaxi's lane-keeping controller, a delivery drone's flight stabilization, an autonomous ship's course control, and a planetary rover's attitude control all classify in the same group, because the underlying problem — controlling position, course, altitude, or attitude — is the same regardless of the medium. The CPC definitional line establishes that scope directly.

Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots— CPC group G05D 1/00 title, source

Below G05D 1/00 the scheme branches into the components of autonomous control. There are groups for controlling a single variable, for simultaneous control of position and course, for two- and three-dimensional control, for adaptations specific to land vehicles, water craft, aircraft, and spacecraft, and for the navigation, obstacle-avoidance, and remote-control mechanisms that autonomous systems depend on. The digits after G05D 1 narrow the subject from "control of a vehicle's course" down to a specific control function — which is how an analyst separates a portfolio's flight-control filings from its ground-vehicle navigation filings, or its obstacle-avoidance claims from its formation-control claims.

How a real G05D claim reads

The class tells you the field; the claim tells you the substance. US9612123B1, "Adaptive mapping to navigate autonomous vehicles responsive to physical environment changes," a granted patent assigned to Zoox Inc, is a G05D-family record directed to keeping autonomous-vehicle maps current. Its claim 1 recites receiving sensor data from "multiple driverless vehicles in a fleet," generating "data sets of three-dimensional map data," "detecting a change between the generated data sets of three-dimensional map data and a stored data set... representative of a state change in the environment," updating the stored map, and "transmitting the updated data set of three-dimensional map data to the multiple driverless vehicles in the fleet," where at least one vehicle "is controlled based at least in part on the updated data set." The claim is directed to the navigation-data side of vehicle control — the maps that the course-control system relies on — which is exactly the kind of subject G05D 1/00 covers.

A different G05D record fences off a different control function. US9229453B1, "Unified motion planner for autonomous driving vehicle in avoiding the moving obstacle" (assigned to GM Global Technology Operations LLC), claims a method that detects a lane center, determines a lane-centering path, automatically steers along it, detects a moving object, predicts a collision, and computes a steering path around the object when braking will not suffice. That is the planning-and-steering side of vehicle control. Both records live in the G05D autonomy neighborhood, but one fences off fleet map-maintenance and the other fences off obstacle-avoidance planning. The shared class symbol does not make their claims overlap; only the limitations do.

Why G05D is the anchor class for autonomy search

The relationship between G05D and the conjoint-vehicle-control class B60W is worth making explicit, because the two are easy to confuse and frequently co-occur. G05D is the general control class — it covers controlling the position, course, altitude, or attitude of a vehicle as an abstract control problem, across land, water, air, and space. B60W, by contrast, is a road-vehicle-specific class whose title is "CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION" — it concerns coordinating a car's subsystems (steering, braking, powertrain) together. An autonomous-driving invention often touches both: the navigation-and-course logic classifies in G05D, while the coordination of the car's actuators to execute that course classifies in B60W. That is why so many self-driving patents carry symbols in both classes, and why an analyst seeding a search on only one of them will systematically miss part of the estate. The two classes are complementary lenses on the same autonomy stack, not substitutes.

For anyone mapping who owns what in autonomous driving, G05D 1/00 is the front door, the way B25J is the front door for manipulators. A search seeded by G05D 1/00 subgroups surfaces the navigation-and-control estate — path planning, sensor-driven course control, fallback logic, fleet coordination — across self-driving, drone, and robotics assignees. Because the CPC is language-independent, a G05D 1/00 filing in Japanese or German carries the same symbol as one in English, which makes cross-jurisdiction autonomy portfolio mapping tractable. Three practical notes. First, autonomy patents frequently carry a G05D symbol alongside a B60W conjoint-vehicle-control symbol and often a perception or machine-learning class; seed searches on G05D and B60W together to avoid undercounting. Second, the digits matter — a portfolio heavy in obstacle-avoidance subgroups but thin in fleet-coordination subgroups tells a different strategic story than the reverse. Third, the authoritative scope of any G05D group is the official CPC scheme published by the USPTO and EPO, not an aggregator summary; when the exact boundary of a group matters, the scheme text is the primary source. G05D 1/00 — control of position, course, altitude, and attitude — is precisely the field where the autonomy-control patent fight is being fought.