MEMS Accelerometer AI/ML Patent Landscape 2026
The intersection of MEMS inertial sensing and machine learning is in sustained growth, with 663 patent families on record and annual filings rising sharply since 2017. Samsung Electronics holds the lead position, but the field remains fragmented across consumer electronics, academia, and industrial players, signalling an open competitive window.
Samsung leads a fragmented field with no dominant bloc
Samsung Electronics Co. Ltd. holds the top position with 24 patent families, followed by Google LLC at 14 and Wuhan University at 13. The top five filers—Samsung, Google, Wuhan University, Lovely Professional University, and Qualcomm—together account for 17% of the combined output of the hundred largest filers, signalling a genuinely fragmented competitive landscape.
No single player commands a decisive share. The gap between first and second place (24 vs. 14 patent families) is meaningful but not entrenching; challengers including Qualcomm, Southeast University, and Honeywell International each hold 9–10 patent families, placing them within striking distance of the top tier.
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 1 | Samsung Electronics Co. Ltd. | 24 | |
| 2 | Google LLC | 14 | |
| 3 | Wuhan University | 13 | |
| 4 | Lovely Professional University | 10 | |
| 5 | Qualcomm Incorporated | 10 | |
| 6 | Southeast University | 10 | |
| 7 | Honeywell International Inc. | 9 | |
| 8 | PROCTER & GAMBLE CO | 8 | |
| 9 | Tsinghua University | 8 | |
| 10 | KOREA UNIV RES & BUSINESS FOUND | 8 |
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 11 | Braun GmbH | 7 | |
| 12 | Chongqing University OF POSTS & TELECOMM | 7 | |
| 13 | Beihang University | 7 | |
| 14 | Meta Platforms Technologies LLC | 6 | |
| 15 | MEI Micro Inc. | 6 | |
| 16 | UNIV OF SCI & TECH OF CHINA | 6 | |
| 17 | Starkey Laboratories Inc. | 6 | |
| 18 | TRX Systems | 6 | |
| 19 | Robert Bosch GmbH | 5 | |
| 20 | Omnibus157 Pty Ltd | 5 |
The presence of Chinese universities—Wuhan, Southeast, Tsinghua, Beihang, and Chongqing University of Posts and Telecommunications—alongside consumer-electronics and industrial incumbents indicates that academic institutions are active innovators and potential licensing or collaboration partners, not merely observers.
The most recent 18–24 months of filings are subject to publication lag and will appear understated until grants and publications catch up; the 2025–2026 data should be treated as a floor, not a ceiling. Longer-window growth, applicant concentration, and technology-route coverage are therefore more reliable signals than the latest-year bar alone.
Rapid growth since 2018, anchored by AI computing and navigation branches
Annual filing volume and the technology branch mix together show a field that is broadening in both pace and application scope, with AI computing as the clear structural backbone and navigation, medical, and vision branches growing in importance.
Annual filing trend
Filings rose from 11 in 2017 to a recorded peak of 152 in 2025, representing 80% growth in the recent window. The steep climb from 2021 onward reflects accelerating adoption of ML inference on inertial data. The 2025–2026 bars are understated due to publication lag and will revise upward as patents publish.
↗ Hover for values · click a bar to ask EurekaTechnology composition
G06N (AI/ML computing models) dominates with 544 records, confirming that algorithmic innovation is the primary driver. G06F (digital data processing, 275 records) and G01C (navigation and gyroscopes, 212 records) are the next largest branches, reflecting strong overlap with positioning and sensor-fusion applications. G01P (velocity and acceleration measurement, 141 records) represents the core MEMS hardware interface, while A61B (medical diagnosis, 113 records) and G06V (image/video recognition, 122 records) signal active application-layer expansion into health and computer vision.
↗ 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.
Inertial sensor error modeling and compensation, a…
A method for inertial sensor error modeling and compensation comprises obtaining multiple bias drift datasets for an elapsed time period for one or more gyroscopes; generating a 3D bias drift data plot using the multiple bias drift datasets; generating a partial bias drift data image based on the 3D bias drift data plot; and inputting the partial bias drift… (excerpt from the patent abstract)
| # | Patent | Citations |
|---|---|---|
| 1 | Program Setting Adjustments Based on Activity Iden… | 202 |
| 2 | Method and apparatus for sensing a rollover | 162 |
| 3 | Eyewear having human activity monitoring device | 108 |
| 4 | Body movement tracking | 85 |
| 5 | Systems and methods for deep localization and segm… | 82 |
| 6 | Detection of physical abuse or neglect using data… | 62 |
| 7 | Systems and methods for formulating a performance … | 61 |
| 8 | 一种面向动态场景的单目视觉惯性SLAM方法 | 53 |
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
The combination of a growth-stage lifecycle, low concentration, active academic participation, and. China-heavy geographic filing creates a distinctive risk-opportunity profile for new entrants and incumbents alike.
Growth stage with sustained upward trajectory
The lifecycle evidence places this field firmly in the Growth stage: annual filings have risen continuously from 11 in 2017 to 152 in 2025, and the 80% recent-window growth rate has not yet eased from its peak. For R&D planners, this means foundational positions are still being established and early movers retain a meaningful advantage over late entrants. Patent thickets are not yet dense enough to block new technical approaches.
Growth stageLow concentration creates entry opportunities
The top five filers hold only 17% of the hundred largest filers’ combined output, and no applicant exceeds 24 patent families. This low Herfindahl profile means no single player can exercise blocking leverage across the full technology space. A focused portfolio of 15–20 well-placed patent families could realistically place a new entrant within the top five.
FragmentedSamsung–Korea University pairing is the standout co-filing relationship
The most active co-applicant pair is Samsung Electronics and Korea University Research & Business Foundation, with 8 jointly filed patent families. Samsung also co-filed once with Beijing Samsung Communications Technology Research. Wuhan University co-filed once with OPPO (Guangdong Mobile Communications). These relationships suggest that university-industry co-development is the primary collaboration model, and that aligning with a research institution can accelerate Samsung-tier output.
University-industryChina is the primary filing jurisdiction; US and India are secondary
China accounts for 343 patent records filed, the United States for 150, and India for 75—together covering the vast majority of activity. WIPO PCT filings (42 records) and EPO (34 records) suggest a subset of filers pursuing international protection, but global coverage remains thin. Applicants seeking protection outside their home market should prioritise PCT and EPO routes given the current gap.
China-ledGo 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 |
|---|---|---|
| Samsung Electronics Co. Ltd. | Korea University Research & Business Foundation | 8 |
| Samsung Electronics Co. Ltd. | Beijing Samsung Communications Technology Research Co. Ltd. | 1 |
| Wuhan University | OPPO (Guangdong Mobile Communications Co. Ltd.) | 1 |
Co-filing pairs, ranked by the number of jointly-filed patent families.
Samsung and Google lead on AI computing; Honeywell and Qualcomm anchor navigation
The top two filers are consumer and platform-technology majors with broad AI/ML portfolios, while mid-tier players carve out distinct niches in navigation, sensor hardware, and medical applications.
Samsung Electronics Co. Ltd.
Samsung holds 24 patent families, the largest position in the ranking, with recent filings (18 in the recent period) classified as a new entrant trajectory—indicating its MEMS AI portfolio is a newly established and rapidly expanding focus area. Technology emphasis is concentrated in G06N3 (AI/neural-network computing, 18 filings), G06F3 (digital data processing, 10), and G06T19 (image data processing, 8), pointing to a strategy centred on on-device inference and augmented-reality interaction. Its co-development relationship with Korea University Research & Business Foundation (8 jointly filed families) reinforces an academic pipeline.
patent families: 24Google LLC
Google holds 14 patent families, also classified as a new entrant in momentum terms, with 12 filed in the recent period. Its technology mix spans G06N3 (AI/neural-network computing, 9), G06F3 (digital data processing, 8), and G01P15 (velocity and acceleration measurement, 5)—the last being notable as it connects AI inference directly to raw accelerometer signal processing, a lower-level technical position than most competitors occupy. This suggests Google is pursuing sensor-model co-optimisation, not solely application-layer AI.
patent families: 14| Applicant | Recent (3 yrs) | Trend |
|---|---|---|
| Samsung Electronics Co. Ltd. | 18 | ▲ new entrant |
| Google LLC | 12 | ▲ new entrant |
| Wuhan University | 6 | ▲ new entrant |
| Qualcomm Incorporated | 6 | ▲ new entrant |
| Lovely Professional University | 7 | ▲ new entrant |
| Southeast University | 4 | ▲ new entrant |
| Honeywell International Inc. | 2 | ▲ new entrant |
| Korea University Research & Business Foundation | 4 | ▲ new entrant |
Under-served branches in velocity measurement, medical sensing, and radar-inertial fusion
Several IPC branches adjacent to the dominant G06N computing core carry meaningful patent activity but remain sparsely populated relative to the overall corpus, suggesting areas where focused work could differentiate a portfolio.
G01P · Velocity & acceleration measurement
With 141 patent records and a 7% share of the corpus, G01P sits at the hardware-algorithm interface: innovations here address how raw accelerometer signals are conditioned, calibrated, or interpreted before being passed to ML models. This branch is technically foundational yet under-populated relative to the dominant AI-computing branch, and entry requires sensor hardware expertise combined with signal-processing IP—a combination most pure software filers lack. Organisations with MEMS design capability are best positioned to build here.
Search this in Eureka →A61B · Medical diagnosis and surgery
A61B carries 113 patent records (5% share), spanning fall detection, rehabilitation monitoring, and surgical guidance—all areas where MEMS accelerometers combined with ML classifiers are increasingly viable alternatives to clinical-grade instruments. The branch is adjacent to the core AI stack but requires regulatory and clinical validation expertise that most electronics filers do not possess, creating a structural gap. Medical-device companies or digital-health startups with existing FDA/CE pathways represent the most realistic entrants.
Search this in Eureka →How leaders differ by technology route across AI, navigation, and hardware branches
Route coverage across the main technology branches in the current evidence set.
| Player | G06N 3 · Computing based on AI models | G01C 21 · Distance, navigation & gyroscopes | G06F 18 · Electric digital data processing | G01P 15 · Velocity & acceleration | A61B 5 · Diagnosis & surgery |
|---|---|---|---|---|---|
| Samsung Electronics Co. Ltd. | Strong · 18 | Absent | Absent | Moderate · 6 | Emerging · 3 |
| Wuhan University | Strong · 12 | Moderate · 4 | Moderate · 6 | Absent | Absent |
| Southeast University | Strong · 10 | Strong · 6 | Emerging · 2 | Absent | Absent |
| Qualcomm Incorporated | Strong · 10 | Strong · 6 | Absent | Absent | Absent |
| Honeywell International Inc. | Strong · 9 | Strong · 7 | Absent | Absent | Absent |
| Beihang University | Strong · 7 | Moderate · 3 | Strong · 5 | Absent | Absent |
| Google LLC | Strong · 9 | Absent | Absent | Strong · 5 | Absent |
Frequently asked questions
The analysed corpus contains 663 patent families. Annual filings have grown from 11 in 2017 to a recorded 152 in 2025, with 2025–2026 figures expected to revise upward as pending applications publish.
Samsung Electronics Co. Ltd. holds the top position with 24 patent families, followed by Google LLC at 14 and Wuhan University at 13. The top five filers together account for 17% of the hundred largest filers’ combined output.
China is the primary filing jurisdiction with 343 patent records, followed by the United States at 150 and India at 75. WIPO PCT (42 records) and EPO (34 records) cover international protection routes, though global coverage remains relatively concentrated in these three primary markets.
G06N (computing based on AI models) is by far the largest branch with 544 patent records, reflecting the centrality of neural-network and ML inference to the field. G06F (digital data processing, 275 records) and G01C (navigation and gyroscopes, 212 records) are the next most active branches.
The most active co-applicant pair is Samsung Electronics and Korea University Research & Business Foundation, with 8 jointly filed patent families. Samsung also co-filed once with Beijing Samsung Communications Technology Research, and Wuhan University co-filed once with OPPO (Guangdong Mobile Communications).
The highest-cited work in the corpus relates to activity identification-based program adjustments (202 citations), rollover detection methods (162 citations), human activity monitoring integrated into eyewear (108 citations), and body movement tracking (85 citations). These citations confirm that activity recognition and inertial context sensing are the dominant application anchors in the field.
Ready to map your own MEMS AI patent landscape?
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.