The ranking of 100 assignees shows a clear leader with 304 records, well ahead of the fifth-ranked company at 46 and the tenth-ranked company at 30. Together the top five hold 43.9% of the 1,935 records in scope, and the top ten hold 53.1%. That means over half the field's patenting activity concentrates in just ten organisations, while the remaining ninety ranked companies split the rest, often with single-digit or single-filing counts.

Filing peaked in 2018 at 271 records and has declined since, with the 2021-to-2024 window showing a 51% drop from 219 to 108 filings. Both those years are past the roughly 18-month publication lag, so the decline is a genuine trend rather than a data artefact. Figures for 2025 and 2026 are still incomplete and should not be read as a continuation of the decline until later filings publish.

G06F (electric digital data processing) and H04L (digital information transmission) are the two largest IPC subclasses, covering 27.8% and 23.8% of the 1,935 records respectively. Core control-and-regulating-systems claims under G05B account for 17.6%, and vehicle-control class B60W for 16.3%. Because a single record can carry multiple IPC classes, these shares add up to more than 100%, and the pattern suggests much of the recent claim-writing sits in data-handling and communications layers rather than in control-loop logic itself.

Sub-areas such as failure-prediction models for field devices, 5G-integrated industrial control networks, and AI-based equipment-state classification show thinner filing density relative to the dominant process-variable and control-loop claim clusters. That does not guarantee those branches are technically easier, but it does mean fewer blocking claims currently sit on them. A freedom-to-operate review of the specific claim language in the ranked leaders' portfolios is still the way to confirm any given branch is genuinely open.

US20190137963A1, filed by Siemens and published 2019-05-09, covers a field device with a self-diagnosis function that flags detected errors, marks its output value as temporarily invalid, and switches to a substitute value using two timers with different expiration windows. The dual-timer design lets the device distinguish a brief transient fault from a persistent one, triggering a safety response only when the fault outlasts the shorter timer. It is a representative record for how process-automation field devices structure safety-oriented fallback behaviour, and its claim scope is a useful reference point when checking overlap in new field-device filings.