Robot Learning & Adaptive Control Patent Landscape 2026
The robot learning and adaptive control space holds 575 patent families and is in a confirmed growth stage, with filings up 48% in the recent window. Mitsubishi Electric Corp leads a moderately concentrated field, with industrial and consumer electronics giants clustered at the top alongside a rising cohort of AI-native robotics startups.
Mitsubishi Electric leads a moderately concentrated field with room for challengers
Mitsubishi Electric Corp holds the top position with 25 patent families, edging LG Electronics Inc (23 patent families) and Mitsubishi Electric Research Laboratories Inc (18 patent families). The top five applicants — which also include Embodied Intelligence Inc and Ocado Innovation Ltd — together account for 23% of the combined output of the hundred largest filers.
That 23% top-five share indicates moderate, not dominant, concentration. No single applicant controls the field, and the gap between rank one and rank five is only 12 patent families, suggesting the leading tier is genuinely competitive rather than entrenched.
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 1 | Mitsubishi Electric Corp | 25 | |
| 2 | LG Electronics Inc | 23 | |
| 3 | Mitsubishi Electric Research Laboratories Inc | 18 | |
| 4 | Embodied Intelligence Inc | 14 | |
| 5 | Ocado Innovation Ltd | 13 | |
| 6 | Samsung Electronics Co Ltd | 11 | |
| 7 | Robotics Research Corp | 10 | |
| 8 | Toshiba Corporation | 9 | |
| 9 | Canon Inc | 8 | |
| 10 | Miso Robotics Inc | 7 |
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 11 | Wootzano Ltd | 7 | |
| 12 | KUKA Deutschland GmbH | 7 | |
| 13 | Fanuc Ltd | 7 | |
| 14 | Mako Surgical Corp | 6 | |
| 15 | Tianjin Saiwei Industrial Technology Co Ltd | 6 | |
| 16 | Her Majesty the Queen as Represented by the Minister | 6 | |
| 17 | Cilag GmbH International | 6 | |
| 18 | Tianjin Enzo Technology Development Co Ltd | 5 | |
| 19 | IBM Corporation | 5 | |
| 20 | Massachusetts Institute of Technology | 5 |
The positions of Mitsubishi Electric’s parent company and its dedicated research laboratory both appearing in the top three points to a deliberate dual-track filing strategy, combining platform-level robot control with more focused AI-model research. Meanwhile, Embodied Intelligence Inc and Ocado Innovation Ltd signal that well-funded startups are already competing directly with established industrial names.
Filings from approximately 2024–2026 are understated due to patent publication lag; the apparent slowdown in the most recent period should not be read as a real shift in activity. Longer-window growth, applicant concentration, and technology-route coverage are therefore more reliable signals than the latest-year bar alone.
Sustained growth with a deep core in manipulator robotics and expanding AI-model coverage
Annual filing activity and technology class distribution together show a field that is growing on a multi-year basis and is anchored in physical robot manipulation, but is broadening into AI inference and surgical applications.
Annual filing trend
Activity was negligible through 2017, then accelerated sharply from 2018 onward, reflecting both the commercialization of collaborative robots and the integration of machine-learning techniques into control loops. The 2025 bar is the highest recorded in the dataset; the 2026 figure is a partial-year count and substantially understated by publication lag — it should not be interpreted as a slowdown.
↗ Hover for values · click a bar to ask EurekaTechnology composition
B25J (Manipulators & robots) is overwhelmingly dominant, confirming that hardware-level robot control remains the primary patent venue. G05B (Control & regulating systems) and G06N (Computing based on AI models) are the next two branches, reflecting the fusion of classical control theory with neural and probabilistic learning methods. A61B (Diagnosis & surgery) ranks fourth, marking medical robotics as a meaningful secondary application domain already attracting dedicated IP.
↗ Hover for values · click a bar to ask EurekaHighly cited patent families surfaced by the query
Citation-heavy patent families returned by the query. Use this section as citation context, not as a curated list of the most topic-specific patents.
Robot arm control device, method for training hier…
A robot arm control device is proposed. The robot arm control device may include a memory. The device may also include an acquisition unit acquiring an arbitrary target image, a virtual canvas image in which a virtual drawing operation of a robot arm for the target image is reflected, and a virtual nib image of the robot arm whose position is changed as the… (excerpt from the patent abstract)
| # | Patent | Citations |
|---|---|---|
| 1 | Method and apparatus for adaptive force and positi… | 328 |
| 2 | Adaptive vision-based controller | 323 |
| 3 | Industrial robot with servo | 312 |
| 4 | Adaptive tracking vision and guidance system | 254 |
| 5 | Sensing arrangements for robot-assisted surgical p… | 235 |
| 6 | Industrial robot with controller | 221 |
| 7 | Motion controller for redundant or nonredundant li… | 175 |
| 8 | Method and apparatus for configuration control of … | 169 |
Ranked by total forward citations. Citation counts favour older and broadly cited patent families, and broad or adjacent patents may appear when they match the search scope. Treat this section as citation context, not as a curated list of the most topic-specific patents. Some patent titles may be shown in their original, non-English language where an accurate translation could not be guaranteed.
What the competitive structure means for R&D investment decisions
Four structural signals — lifecycle stage, applicant concentration, collaborative filing activity, and geographic spread — each carry distinct implications for where and how to position new R&D.
Growth stage: filings are still rising on a multi-year basis
The evidence classifies this field as Growth, with annual filings still rising and the most recent years understated by publication lag. A 48% recent-window growth rate confirms meaningful acceleration rather than a mature plateau. Entrants who file now can still shape the prior-art landscape rather than navigating around an already-dense corpus.
Growth stageTop-five hold 23% of the largest-filer pool — no single gatekeeper
The top five applicants account for 23% of the combined output of the hundred largest filers, a moderate share that leaves the field open to challengers. The narrow gap between rank one (25 patent families) and rank five (13 patent families) means that a focused two-to-three-year filing program could move a new entrant into the visible competitive tier. However, the presence of well-resourced industrial incumbents and funded startups simultaneously means quality of claims matters as much as volume.
Moderate concentrationCo-filing activity is sparse, with one material partnership identified
The most active co-filing relationship in the evidence is between Kindred Systems Inc (US) and Ocado Innovation Ltd, with four joint filings recorded — notable given that Ocado ranks fifth overall. A second co-filing pair involves Robotics Res and Roboteitsukusu Res Corp. The overall scarcity of collaborative filings suggests the field is still largely prosecuted by individual applicants, which may reflect competitive sensitivity around core learning algorithms. Organizations seeking differentiated positioning could explore joint filings in application-specific sub-domains where neither party holds a strong solo position.
Limited co-filingUS leads filing volume; China and India are substantial secondary venues
The United States is the lead jurisdiction with 232 patent records, followed by China with 157 and India with 82 — a distribution that reflects both the location of key assignees and the strategic importance of those markets. WIPO (PCT) filings number 63 and European Patent Office filings 56, indicating that the leading applicants are pursuing multi-jurisdictional coverage. Japan, the home of several top-ranked assignees, shows only 23 records, which may indicate a preference for PCT over direct Japanese national filings at this stage.
US-China-India triadGo beyond the landscape: Eureka’s TRIZ Solution agent breaks down an R&D problem and returns patented concept solutions, each with a technical approach and cited patent & literature evidence.
| Applicant | Collaborator | Co-filings |
|---|---|---|
| Kindred Systems Inc | Ocado Innovation Ltd | 4 |
| Robotics Research Corp | ROBOTEITSUKUSU RES CORP | 1 |
Co-filing pairs, ranked by the number of jointly-filed patent families.
Mitsubishi Electric vs. LG Electronics: industrial control depth versus full-stack manipulator breadth
The two top-ranked applicants differ in technology emphasis despite sharing the same dominant IPC class, and both are classified as new entrants in the recent filing window — signaling that their current positions were built quickly.
Mitsubishi Electric Corp
Mitsubishi Electric Corp leads with 25 patent families. Its technology focus spans B25J 9 (manipulator kinematics and programming, 25 filings), G06N 3 (neural-network-based AI models, 7 filings), and B25J 13 (robot sensing and control, 3 filings) — a profile that bridges classical robot programming with learned control. The momentum data classifies the company as a new entrant in the recent period, meaning its leadership position was assembled in a concentrated filing burst rather than accumulated gradually. Its affiliated research laboratory, Mitsubishi Electric Research Laboratories Inc, adds 18 patent families with a complementary but more hardware-focused emphasis (B25J 9 and B25J 13), effectively doubling the group’s coverage.
patent families: 25LG Electronics Inc
LG Electronics Inc holds 23 patent families, essentially tied with the leader. Its filing emphasis is concentrated across three B25J sub-classes — B25J 9 (23 filings), B25J 11 (special-purpose robots, 8 filings), and B25J 13 (sensing and control, 7 filings) — with no material presence in G06N, suggesting a hardware-centric strategy focused on robot platforms rather than learning algorithms per se. Like Mitsubishi Electric, LG is classified as a new entrant in the recent window, confirming rapid recent accumulation. Samsung Electronics Co Ltd (11 patent families, also a new entrant) represents a third Korean consumer-electronics player entering the manipulator space with a similar B25J-heavy profile.
patent families: 23| Applicant | Recent (3 yrs) | Trend |
|---|---|---|
| Mitsubishi Electric Corp | 12 | ▲ new entrant |
| Mitsubishi Electric Research Laboratories Inc | 16 | ▲ new entrant |
| Embodied Intelligence Inc | 2 | ▼ -83% |
| Samsung Electronics Co Ltd | 8 | ▲ new entrant |
| Kindred Systems Inc | 2 | ▲ new entrant |
Under-served adjacent branches worth monitoring
Several IPC classes appear alongside the dominant B25J core but carry notably lower filing shares, indicating they are adjacent to current activity rather than saturated — some may offer technically credible entry points.
G05D · Control of non-electric variables
G05D covers autonomous control of physical variables such as position, velocity, temperature, and flow — directly relevant to adaptive control loops in mobile and field robots. It accounts for only 3% of the branch distribution despite being technically adjacent to both G05B and B25J. For teams working on adaptive control in unstructured environments (agricultural, construction, or field-service robotics), this branch remains lightly contested. An entry path exists through filing on sensor-fusion-driven adaptive regulators that fall outside the conventional manipulator framing of B25J.
Search this in Eureka →G06T · Image data processing & generation
G06T (image data processing and generation) carries a 2% share of the branch distribution, despite vision-based feedback being a foundational component of modern learned robot control. The sparse coverage relative to the volume of active vision-guided robot systems suggests that applicants are preferring to file visual-feedback innovations under B25J or G06N rather than G06T directly. Teams with proprietary visual representation learning or novel image-space control architectures may find this branch relatively open. Entry is technically plausible for organizations already active in G06N who can extend claims into the image-processing layer.
Search this in Eureka →How leaders diverge across technology routes
Route coverage across the main technology branches in the current evidence set.
| Player | B25J 9 · Manipulators & robots | B25J 13 · Manipulators & robots | G06N 3 · Computing based on AI models | G05B 19 · Control & regulating systems | B25J 19 · Manipulators & robots |
|---|---|---|---|---|---|
| LG Electronics Inc | Strong · 23 | Moderate · 7 | Emerging · 2 | Absent | Moderate · 5 |
| Mitsubishi Electric Corp | Strong · 25 | Absent | Moderate · 7 | Absent | Absent |
| Mitsubishi Electric Research Laboratories Inc | Strong · 18 | Moderate · 5 | Absent | Absent | Absent |
| Samsung Electronics Co Ltd | Strong · 11 | Strong · 6 | Absent | Absent | Moderate · 5 |
| Embodied Intelligence Inc | Strong · 13 | Absent | Strong · 8 | Absent | Absent |
Frequently asked questions
The evidence covers 575 patent families in scope. The United States is the lead filing jurisdiction with 232 patent records, followed by China with 157 and India with 82.
Mitsubishi Electric Corp leads with 25 patent families. LG Electronics Inc is effectively tied at 23 patent families, and Mitsubishi Electric Research Laboratories Inc holds third place with 18 patent families.
The evidence classifies the field as Growth. Annual filing volume is still rising on a multi-year basis, with a 48% recent-window growth rate. The most recent one to two years of data are understated by publication lag, so the apparent plateau in 2025–2026 does not reflect a real slowdown.
B25J (Manipulators & robots) is overwhelmingly dominant. G05B (Control & regulating systems) and G06N (Computing based on AI models) are the next largest branches, followed by A61B (Diagnosis & surgery) and G06F (Electric digital data processing).
Co-filing activity is sparse. The most substantive partnership identified is between Kindred Systems Inc (US) and Ocado Innovation Ltd, with four joint filings recorded. A second pair — Robotics Res and Roboteitsukusu Res Corp — accounts for one joint filing. The overall low frequency of co-filings suggests most applicants are filing independently.
G05D (Control of non-electric variables) and G06T (Image data processing & generation) each hold a low share of the branch distribution — 3% and 2% respectively — despite being technically proximate to core adaptive control and vision-guided robot applications. B65G (Conveying & material handling) is another branch with relatively limited coverage given its practical relevance to adaptive robot systems.
Map your own position in robot learning and adaptive control
Join 18,000+ innovators using PatSnap Eureka to map any technology landscape: search 2B+ patents and papers, surface key assignees, and generate a report like this in minutes.
Disclaimer. This page is generated from PatSnap Eureka data drawn from a limited snapshot of global patent and scientific-literature records, and is provided for general information and reference only.
Patent data carries inherent limitations: recent filings (typically the most recent 18–24 months) are under-counted due to standard publication lag; counts may be reported at either a patent-family or a patent-record basis and are not always directly comparable; classification, applicant-name, and citation data may contain errors, duplicates, or omissions; and the underlying search query defines and constrains the scope shown. As a result, the analysis may be incomplete or inaccurate and may not reflect the full technology landscape.
Nothing on this page constitutes an exhaustive prior-art, novelty, freedom-to-operate, or validity search, nor does it constitute legal, financial, investment, or professional advice, and it should not be relied upon as such. Any patent, commercial, or strategic decision should be verified independently and reviewed with qualified patent, legal, and domain professionals. PatSnap makes no warranties, express or implied, as to the accuracy, completeness, or fitness for any particular purpose of the information presented.