The dataset tracks assignees including Intuitive Surgical, Johns Hopkins University, Sony Group and Covidien LP among others, but recent-year momentum for all of them is at or near zero filings. Historical citation leadership sits with an early microsurgical robot system patent from the mid-2000s, US7155316B2, which still carries the highest citation count in the corpus. That gap between historical filing depth and current activity suggests the leaders established their positions years ago and have since slowed. Anyone assessing freedom-to-operate should weight older granted patents from these assignees more heavily than recent filings.

Filings rose from 18 in 2017 to a peak of 77 in 2022, and have not returned to that level since. Because publication typically lags filing by around 18 months, the most recent years in any such trend are always undercounted, so the true 2025-2026 picture will revise upward somewhat. Even with that adjustment, the shape looks like a plateau rather than continued growth. Treat this as a maturing filing landscape, not a shrinking one, until a later data pull confirms the direction.

The bulk of records sit in A61B (diagnosis and surgery, 266 of 435) and G09B (educational and demonstration aids, 147), with G16H (healthcare informatics, 106) and G06F (digital data processing, 76) also carrying meaningful volume. G06T (image processing) and G06N (AI models) are present but thinner, at 58 and 31 respectively. That spread shows claims split fairly evenly between the physical surgical robot, the training/simulation apparatus, and the software layer that connects them.

That filing, from Intuitive Surgical Operations, claims a specific combination: a physical tissue-and-skeleton rig, optical tracking of markers to build a 3D wireframe, and force/torque sensing tied to an actual robotic surgery station's patient support. A simulator that uses only software-based tissue modeling without a physical rig, or one not coupled to a live robotic station, sits outside that specific combination. Designers should check whether their system reproduces all three elements together — the rig, the optical tracking, and the station-coupled force sensing — since removing any one of them is a plausible design-around path worth verifying with counsel.

IPC composition shows AI-driven procedural elements (G06N, 31 records) and image processing (G06T, 58 records) are thin relative to the A61B and G09B core. Sub-areas such as haptic tissue-deformation feedback loops, AI-based skill scoring, and cross-platform data portability between simulators and live robots show low filing density in this corpus. Co-assignee activity is also sparse — only 10 pairs across 435 families — meaning collaborative claim territory between companies and universities is largely open. These are reasonable areas to file a first claim before density builds up, as it has in the core surgical-device space.