Filing is concentrated at the top of the ranking: the leading assignee holds 147 records, and the top 5 assignees combined hold 488 records, or 18.5% of the 2,638 records in scope. The top 10 combined reach 699 records, 26.5% of the field, after which the ranking spreads into a long tail of single- and low-volume filers. This pattern is typical of an automotive safety feature where major OEMs and Tier 1 suppliers each maintain their own control-logic and sensor-integration filings rather than one company dominating outright.

Filings peaked at 197 in 2020, then declined to 150 in 2021 and 121 in 2024, a 19% drop over that three-year span. That said, publication lags filing by roughly 18 months, so the lower counts recorded for 2025 and 2026 in the dataset reflect incomplete publication rather than a confirmed slowdown. The safest read is that the field grew rapidly through 2020 and has since settled into a lower but still active filing rate, based on 2024 as the most recent complete year.

The largest classes are B60W (hybrid/joint vehicle control, 31.0% of 2,638 records) and G08G (traffic control systems, 27.3%), both ahead of B60R (vehicle parts and accessories, 20.9%) and B60Q (vehicle lighting and signalling, 16.1%). Sensing and imaging classes — G01S (radar/sonar/positioning, 14.7%), H04N (pictorial communication, 9.2%) and G06V (image/video recognition, 7.8%) — appear less often, indicating that control-layer integration is claimed more heavily than the underlying detection hardware. Because a single record can carry several classes, these shares add up to more than 100% and should not be treated as mutually exclusive categories.

US20140002657A1, filed by LG Innotek and dated 2014-01-02, claims a specific pipeline: a forward-facing camera captures image data, a driving unit classifies candidate vehicles from that data using a predetermined mask, filters candidates to confirm a real object, tracks the object across multiple frames to recover any missed detections, and calculates a collision time from object-to-vehicle distance to generate a warning. It blocks that particular sequence of steps rather than the general concept of forward collision warning, so implementations using a different sensor modality or omitting the multi-frame recovery step may sit outside its literal claims. Anyone building a vision-based warning system should read the full claim set rather than relying on the abstract alone.

The technology composition data points to imaging and recognition classes — G06V at 7.8% and H04N at 9.2% of the 2,638 records — as comparatively under-claimed relative to the dominant control classes B60W (31.0%) and G08G (27.3%). Only 10 co-assignee pairs are documented across the dataset, suggesting joint OEM-supplier filings remain rare and that space is open too. A practical starting point is a claim that ties a specific recognition-confidence or fusion metric to a graduated warning or propulsion response, rather than repeating the classification-then-alert structure that dominates the existing filings.