In this dataset, a minimal risk maneuver refers to a control routine that brings an autonomous vehicle to a safe state when normal operation cannot continue, such as a sensor failure, a system fault, or an unresolvable driving scenario. Filings describe this using varied terminology, including automated fallback maneuver, safe stop maneuver, and autonomous risk response, which is why the search string combines all four phrasings. The core technical content typically covers trigger detection, a planning step that selects a safe stopping location or action, and an execution step that hands control to braking or steering actuators.

The ranked list covers 74 companies, with one filer holding a count roughly eight times that of the fifth-placed company, indicating an early and broad claim position. Behind that leader, filing counts drop off quickly through a competitive mid-tier of automakers and at least one academic research partner that co-files with two of the automakers. Because the ranking is based on 655 total records and not a top-50 or top-100 cut, smaller filers with a handful of records each still make up a meaningful share of the list.

Filing activity grew sharply between 2021 and 2024, rising from 50 to 116 filings, a +132% increase over that three-year span, which is the last period that can be treated as a complete filing-year comparison. Counts for 2025 and 2026 appear lower in raw form, but that reflects the roughly 18-month lag between filing and publication rather than an actual decline. Any statement that the field is slowing should be checked against complete years only.

Vehicle-control claims under IPC class B60W appear on 65.8% of the 655 records in scope, making execution-layer control the most crowded area by a wide margin. Perception, data processing, traffic control, and braking classes each appear on well under 16% of records, and the trigger-and-decision logic that determines when a fallback maneuver should activate is comparatively thin. That gap suggests more room to differentiate claims around detection and decision criteria than around the mechanics of executing a stop or lane change.

A representative filing from Hyundai, published as US20250256704A1, illustrates the pattern: a sensor set generating surrounding-environment information, a processor that monitors vehicle state and controls autonomous driving, and a controller that executes the maneuver under the processor's direction, with detection of the need for a minimal risk maneuver tied to real-time sensed and state data. Claims built this way tend to bundle the trigger condition, the decision logic, and the execution path into a single system claim, which is one reason B60W vehicle-control language appears so consistently across the dataset. Designing around such a claim typically means altering how the trigger condition is derived or how control authority is handed off, rather than the underlying sensors themselves.