AGV AI & Machine Learning Patents: Who Leads, Where Gaps Are 2026
- Filing peaked in 2025 at 33 records after a flat middle stretch (13 in 2022), suggesting the field surged recently rather than growing steadily since 2017.
- China accounts for 96 of 168 receiving-office filings more than five times the United States total of 29, concentrating both prior art risk and freedom-to-operate work in Chinese filings.
- G06N and G05D each cover roughly half the corpus (90 and 89 of 168 records) showing AI-model claims and non-electric control claims are filed almost as pairs, not as separate technology bets.
Filing growth compares 2021 (21 records) with 2024 (20) — a three-year span. 2024 is the most recent year we treat as complete: publication lags filing by roughly 18 months, so 2025 onwards are still filling in and any growth rate that ends there would understate the field. Top-5 share is the combined record count of the five largest assignees divided by all 168 records in scope (CR5), not by the ranked leaders only.
What this patent set covers
This landscape isolates patent families at the intersection of automated guided vehicles and machine learning methods — reinforcement learning path planning, deep learning perception, and predictive navigation — rather than AGV hardware in general. The search combines AGV-specific terminology with AI/ML claim language and IPC codes for neural-network computing (G06N3), non-electric variable control (G05D1/02) and image or video recognition (G06V20), so mechanical-only AGV filings and generic industrial-robot AI filings both fall outside the set.
The 168 records span 2015 through mid-2026, with the most recent year necessarily undercounted because publication typically lags filing by around 18 months. Reading the trend line as a leading indicator of interest, not a finished count, is the safer approach for the 2025-2026 window in particular.
Filing trend and technology composition
Two views of the same 168-family corpus: how filing activity moved year over year, and which IPC subclasses carry the claim weight.
A recent surge, not a steady climb
Filings sat at zero in 2017, reached a midpoint of 13 in 2022, and climbed to a peak of 33 in 2025 before the partial 2026 count of 9. That shape — flat, then a late jump — points to AGV-specific AI/ML claiming becoming an active filing strategy only in the last few years, rather than a technology that has been building steadily for a decade.
AI-model claims and control claims are filed together
G06N (AI models) and G05D (non-electric control) each appear in roughly half the corpus, with G06Q (business/commerce data processing) close behind at 79 — an unusually high share that suggests many filings pair navigation AI with scheduling or fleet-management logic. Perception-specific codes (G06V, G06K) and positioning codes (G01C) trail well behind, each under 20 records, marking those as comparatively open claim territory.
Shares are the percentage of the 168 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Automated Guided Vehicle AI and Machine Learning with Eureka
This page is one run against one query. Ask Eureka your own question about automated guided vehicle ai and machine learning and every answer comes back with the patent numbers behind it.
Try EurekaThe most-cited records in this set
WO2026046502A1 — Method and system for testing and retraining an AGV using synthetic obstacles
A method and system for retraining an Automated Guided Vehicle (AGV) in a factory environment is disclosed. The method involves receiving a visual feed from at least one sensor configured to capture visual data of the factory environment. A plurality of synthetic obstacles is inserted into the received visual feed to generate a modified visual feed, simulating real-world conditions by embedding digital representations of real-world obstacles. The modified visual feed is analyzed using a reinforcement learning-based navigation model to generate a navigational decision based on the detection of the synthetic obstacles.Filed by Siemens Aktiengesellschaft, published 2026-03-05 — illustrates how incumbents are now claiming the training-data pipeline around AGV reinforcement learning, not just the navigation model itself.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | CN113485380A | 一种基于强化学习的AGV路径规划方法及系统 | 65 |
| 2 | KR102043143B1 | Method and apparatus for driving control of automated guided vehicle by using artificial neural network | 50 |
| 3 | CN110443412A | 动态优化加工过程中物流调度及路径规划的强化学习方法 | 44 |
| 4 | US20220317695A1 | Multi-AGV motion planning method, device and system | 28 |
| 5 | CN117213497A | 基于深度强化学习的AGV全局路径规划方法 | 24 |
| 6 | KR102043142B1 | Method and apparatus for learning artificial neural network for driving control of automated guided vehicle | 22 |
| 7 | US20200125109A1 | Autonomous ground vehicle (AGV) cart for item distribution | 19 |
| 8 | CN115933641A | 基于模型预测控制指导深度强化学习的AGV路径规划方法 | 18 |
| 9 | CN112015174A | 一种多AGV运动规划方法、装置和系统 | 18 |
| 10 | CN116307464A | 一种基于多智能体深度强化学习的AGV任务分配方法 | 17 |
Citation counts reflect influence within the searched corpus and skew toward older filings; treat them as a signal of who shaped the field's vocabulary, not of which claims are strongest today.
Patent titles are shown in the language they were filed in, not translated, so that each record stays verifiable against the original filing — a translated title will not match in Eureka or in any national register. Each row carries its publication number; clicking a row searches Eureka by that number.
Put your own technology through the same analysis
Eureka on the web
When you want the answer in the next five minutes.
The agent works the prompt against patents and technical literature, citing every source.
Run your analysis now →MCP server & REST API
When it has to run inside your own pipeline.
Patent search, landscape analysis and assignee resolution as MCP tools. Drop them into any agent framework, or call REST directly.
Browse MCP servers →What the numbers mean for filing strategy
Three patterns in the data carry direct implications for where to file, what to cite, and what to watch.
China is the primary prior-art jurisdiction
With 96 of 168 records filed through China against 29 through the United States, any freedom-to-operate search that skips Chinese-language filings is working with a fraction of the relevant art. The gap is wide enough that it is a structural feature of this field, not noise.
AI-model claims rarely stand alone
G06N (AI models) and G05D (non-electric control) sit almost level at 90 and 89 records, meaning most filers are claiming the learning model and the control-loop implementation together rather than protecting the algorithm in isolation.
Reinforcement learning path planning set the vocabulary
The most-cited record and several others in the top tier center on reinforcement-learning-based AGV path planning, indicating this sub-area supplied much of the field's foundational claim language that later filings had to work around or build on.
The surge is recent and narrow
Growth from zero in 2017 to a peak of 33 in 2025, after a flat midpoint of 13 in 2022, shows this is a young filing wave. Several major assignees show zero filings in the latest tracked year, so leadership at the top is not yet locked in.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to automated guided vehicle ai and machine learning, with the prior art for and against each one.
Who is filing, and where the field is still open
The assignee list mixes established industrial names with universities and research institutes, and recent-year momentum is muted across the board — a sign the leaderboard is still being contested rather than settled.
No assignee is pulling away
Multiple tracked assignees, including large industrial names and research institutes, show zero filings in the most recent tracked year, and at least one shows a -100% year-over-year change. That pattern is consistent with a field where the leaderboard reflects historical filing rather than current-year activity.
Co-filing is limited and individual-led
The strongest co-assignee pair in the dataset is two named individuals rather than two companies, and the pair count is small. Formal cross-company collaboration on AGV AI/ML claims does not yet show up as a dominant pattern in this corpus.
US filers are a minority presence
United States receiving-office filings sit at 29 against China's 96, roughly a third of the volume. Assignees building a global AGV AI strategy should expect the bulk of comparable art, and of future competitors, to surface first in Chinese filings.
| Assignee | Recent year | YoY |
|---|---|---|
| Walmart Apollo, LLC | 0 | — |
| LingDong Technology (Anhui) Co., Ltd. | 0 | — |
| Siemens AG | 0 | -100% |
| Hangzhou Dianzi University | 0 | — |
| Industrial Technology Research Institute (ITRI) | 0 | — |
| Vibrolo Co., Ltd. | 0 | — |
| GoerTek Inc. | 0 | — |
| GARIMA SINGH | 0 | — |
Working with this landscape
The dataset points to open sub-areas and a jurisdiction imbalance that are worth checking against a specific claim set before drafting or clearing.
Check freedom-to-operate against Chinese filings first
Given the 96-of-168 concentration in China's receiving office, a clearance search that starts elsewhere is starting from the smaller pool.
Run a Eureka searchDraft around the perception and positioning gap
G06V, G06K and G01C each cover fewer than 20 records, well below the G06N/G05D pairing that dominates the corpus — a narrower field to search before claiming there.
Explore white space in EurekaCommon questions about AGV AI and ML patents
This dataset tracks 168 patent families published between 2015 and mid-2026 that combine automated guided vehicle terminology with AI/ML claim language such as reinforcement learning, deep learning perception, or predictive navigation. That figure is not the full universe of AGV patents generally, since it excludes filings that are purely mechanical or that use AI terms without AGV-specific claims. It also understates the most recent year or two, because publication typically lags filing by around 18 months, so 2025-2026 counts will keep rising as more applications publish.
The corpus shows filing concentrated among a mix of industrial automation firms, e-commerce/logistics companies, and Chinese universities and research institutes, but no single assignee dominates recent-year filing: several named organizations, including large industrial players, recorded zero filings in the latest tracked year. The strongest co-filing relationship in the data is between two named individual inventors rather than two companies, which suggests collaborative or corporate-partnership filing is still limited in this niche. Anyone naming a single 'leader' should check the year of that claim, since rankings by cumulative count can differ sharply from current-year momentum.
China's receiving office accounts for 96 of the 168 records in this dataset, more than three times the United States total of 29, and well ahead of WIPO PCT, India, Europe and South Korea combined. This reflects both the concentration of AGV manufacturing and warehouse-automation deployment in China and a general pattern of domestic-first filing before international phase entry. For competitive intelligence or clearance work, this means Chinese-language filings carry a disproportionate share of the relevant prior art and should not be treated as a secondary search after English-language sources.
IPC codes tied to perception and positioning, specifically image/video recognition (G06V), data recognition (G06K) and navigation/gyroscope codes (G01C), each appear in fewer than 20 of the 168 records, well below the roughly 90-record levels for AI-model claims (G06N) and non-electric control claims (G05D). That gap suggests multi-AGV perception fusion, degraded-sensor navigation, and positioning-integration claims are comparatively under-claimed relative to core path-planning and control filings. Under-claimed does not mean unclaimed, so any drafting into these areas still needs a targeted search rather than an assumption of open space.
A high citation count inside a searched corpus, such as the 65 citations recorded for CN113485380A, indicates the record has been referenced often by later filings and likely shaped claim language others had to work around. It is a signal of historical influence, not a guarantee of current enforceability or scope, and citation counts structurally favor older records simply because they have had more time to be cited. Anyone assessing risk from a highly cited record should read its actual claims and current legal status rather than relying on the citation count alone.
Research Automated Guided Vehicle AI and Machine Learning in depth with Eureka
Go past this page: query the whole automated guided vehicle ai and machine learning corpus yourself, in your own scope.
Every answer comes back with patent numbers you can open.
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.
Machine translation. Assignee and organisation names originally recorded in Chinese, Japanese or Korean have been rendered into English by an AI translation step so that the tables stay readable. These renderings are best-effort and may not match a company’s registered English name; the original name is what the underlying patent record carries, and it is what any Eureka query launched from this page uses.