An initiative of AutoWrap Robotics + AutoWrap AI
SafeHumanoids

Humans First. American Innovation.

Building a safer future for humanoid robotics in America.

By Raoul Shah

4
Issued U.S. patents
45
Granted claims across the robotics set
3
Registered U.S. trademarks
2025–26
Grant dates
Robot-agnostic by claim — targets named in the specifications and integration layer
HumanoidArticulated armGantryMobile platformROS 2 · Nav2 · MoveIt
The initiative

Safe humanoid robotics in America.

SafeHumanoids is an initiative of AutoWrap Robotics and AutoWrap AI dedicated to showcasing and advancing the development of safe humanoid robotics in America. The initiative focuses on human-robot interaction, environmental awareness, intelligent safety, validation, and responsible deployment of humanoid systems alongside people.

Mission

Advance American innovation in humanoid robotics by developing and showcasing technologies that enable intelligent machines to operate more safely and responsibly alongside people.

American invention and engineering

The work behind the initiative is American-invented and American-engineered: four issued United States patents, and the control software written against them.

Human-centered by design

Supervised autonomy, bounded work zones, human-presence detection and an operator who can stop the machine at any moment — enforced in the software, not just described.

Responsible development

What is built and what is not is stated plainly, including the stages that have not been completed on production hardware.

How this fits together

One ecosystem, three parts.

SafeHumanoids is not a separate company. It is a dedicated initiative within the AutoWrap technology ecosystem, created to focus attention on the safety question that the humanoid era turns on.

Robotics & physical automation
AutoWrap Robotics

The issued patent portfolio and the physical automation it describes — surface treatment, vehicle lighting, service centre orchestration. autowraprobotics.com →

AI & software intelligence
AutoWrap AI

The control platform: perception, planning, execution and inspection, built to run on robots it does not have to own. autowrap.ai →

This initiative
SafeHumanoids

Safe humanoid robotics in America — the technologies, engineering and safety systems needed for humanoid robots to work responsibly alongside people. By Raoul Shah.

About

Why it exists.

SafeHumanoids was created by Raoul Shah as an initiative within the AutoWrap Robotics and AutoWrap AI ecosystem to focus attention on one of the defining challenges of the humanoid era: enabling intelligent machines to operate safely alongside people. America is entering the humanoid robotics era, and the question of how these machines behave around people is being decided now, in the engineering.

The portfolio

Four issued U.S. patents. Verifiable on the public record.

Every number below can be checked against USPTO or Google Patents. All four are granted — not applications, not provisionals, not pending.

PatentTitleFiledIssuedClaims
U.S. 12,643,243 B1App. 19/260,568 AI platform for automated vehicle surface treatment using humanoid robots Robotic application of paint protection film, vinyl wrap, ceramic coating, window tint and graphics — to vehicles, watercraft, and building window surfaces. 6 Jul 20252 Jun 202610 · system
U.S. 12,649,240 B1App. 19/266,236 AI-led platform for automated vehicle treatment using humanoid robots Adaptive robotic installation and servicing of vehicle lighting — headlights, tail lights, fibre-optic headliners and underbody strips — in confined, variable conditions. 11 Jul 20259 Jun 202615 · system + method
U.S. 12,511,598 B1 Humanoid AI for autonomous vehicle service centres A central AI orchestrating a humanoid fleet across a facility: explainable diagnostics for ICE, EV and software-defined vehicles, task re-allocation, digital twins, geofenced zones. Dependent claims reach high-voltage EV handling, and humanoids that drive a vehicle into a bay and schedule charging against grid demand and electricity pricing. 10 Jul 202530 Dec 202520 · system, method, CRM
U.S. 12,507,946 B1 AI-driven dermatological imaging and predictive simulation platform An adjacent, separable asset applying the same imaging-to-forecast lineage outside mobility. Available for licence independently of the robotics set. 23 Aug 202530 Dec 2025method
Assignee of record
AutoWrap Robotics LLC
Named on the face of all four patents
Inventor
Raoul Shah
West Palm Beach, Florida — address of record at filing
Now based in Virginia
Registered U.S. trademarks
AUTOWRAP ROBOTICS® · AUTOWRAP AI® · FADE MY INK®

Scope. Nothing on this site is legal advice — counterparties should have their own IP counsel review claim scope before any agreement.

The software

The claims, reduced to working software.

The portfolio is not a paper position. Each patent has a corresponding control platform you can run right now in a browser — the orchestration logic, visualised. Open one and watch it work.

00 — Full operation

Operations Control Tower

The whole autonomous operation in one view: a central AI orchestrating a fleet of arms, gantries, mobile units and humanoids through the Scan → Plan → Apply → quality control → corrective-action pipeline, across vehicles, marine hulls and building glazing — with live digital-twin metrics, geofenced zones, and automatic task re-allocation when a unit drops offline.

Spans the three robotics patents · 12,511,598 · 12,643,243 · 12,649,240
01 — Fleet orchestration

Autonomous Service Centers

A central AI coordinates a fleet of humanoids to run a whole service center — intake, Explainable-AI diagnostics (including EV/SDV), repair, quality control, and customer handoff.

U.S. Patent 12,511,598 B1
02 — Surface treatment

Autonomous Surface Treatment

Robots apply paint protection film, ceramic coating, vinyl wrap, tint, and graphics via real-time 3D surface mapping, automated multi-tool switching, machine-vision defect detection, and fleet learning.

U.S. Patent 12,643,243 B1
03 — Adaptive control

Vehicle Lighting Treatment

A humanoid services vehicle lighting in tight cabins, dynamically adapting its posture, motion path, and care to sensed conditions — remove, install, align, restore, and inspect.

U.S. Patent 12,649,240 B1
Licensing

How companies work with us.

We are a licensor. The portfolio and the software behind it are available to companies who want to apply them in their own products, lines or service networks — structured to fit how you actually operate. Structures below are starting points for discussion, not an offer of terms.

Structure 01

Field-of-use licence

A licence to practise the claims within a defined application, territory and term — the usual route where you already have manufacturing or service operations to apply it to.

  • Scoped by application, not the whole portfolio
  • Exclusive or non-exclusive within the field
  • Running royalty or fixed-fee structures
Structure 02

Joint development

We adapt the control platform to your robot, your material and your process, with the resulting work product and any improvements allocated up front.

  • Integration onto your chosen robot platform
  • Process tuning for your film, coating or component
  • Defined ownership of improvements and new filings
Structure 03

Multi-field or portfolio licence

A licence spanning several fields of use, or the portfolio as a whole, for a licensee whose operations cross more than one application. Outright assignment of the portfolio is also available — see licensing & acquisition.

  • Several fields under one agreement
  • Software and simulation assets included
  • Direct support from the named inventor
Issued claims, not applications

Four granted U.S. patents with verifiable numbers, filing dates and grant dates — the position exists today.

Working control software

The orchestration logic runs now: browser platforms and a ROS 2 / Gazebo simulation package, not slideware.

Robot-agnostic integration

The claims recite humanoid, articulated-arm, gantry and mobile platforms. Swapping the robot is configuration, not a rewrite.

Recorded evidence

Thirty-five recorded simulation runs, each captioned with the claim element it exercises. Watch them →

Direct inventor support

Raoul Shah is the named inventor on all four patents and works with licensees directly — on site in the United States for integration and process tuning, and remotely thereafter. Scope and duration are set per agreement.

How it works

One AI brain. Any robot. Safety at every step.

The control platform runs in simulation today, behind a hardware-abstraction layer written so the same behaviour can drive a physical robot without rewriting it. It has not yet driven one. Validating in simulation before anything moves a motor is the point of the layer, and the physical stage is still ahead.

Sense & model

Cameras and depth sensors build a real-time 3D model of the vehicle and workspace.

Plan & adapt

The AI plans the task and adapts posture, motion, and parameters to conditions.

Act, supervised

The robot executes within safety-rated limits, monitored, with instant human stop.

Verify & log

Machine-vision QC verifies the work; every decision is explainable and audit-logged.

Safety, compliance & trust

The SafeHumanoids Framework.

Autonomy near people and property demands more than capability — it demands trust. Every platform is engineered to a safety framework aligned with recognized U.S. and international standards, and enforced in the software itself.

Hardware safety & E-stop

Emergency stop, safety-rated monitored stop, and force/speed limiting. The robot stops on any fault, first.

Supervised autonomy

Human-in-the-loop by default. Operators can pause, override, or stop at any moment.

Geofencing & presence

Work zones are bounded; human-presence detection slows or halts motion in the cell.

Explainable & auditable AI

Every diagnosis and action traces to its inputs and is recorded in a tamper-evident log.

Cybersecurity & privacy

RBAC, signed updates, encrypted telemetry, real-time PII anonymization, and U.S. data handling.

Built for the USA

Designed for American shops and workers, with a compliance roadmap to U.S. safety requirements.

ISO 10218ISO/TS 15066ANSI/RIA R15.06ISO 13849NIST AI RMFOSHAUL — certification path

In plain terms: SafeHumanoids is early-stage. These standards define the framework our systems are engineered to and our active compliance roadmap — they are design targets and a path to certification, not claims of completed third-party certification. We publish this because trust starts with transparency.

Maturity

What is built, and what is not.

Stated plainly, because a licensee's engineers will establish it in the first technical call anyway.

Available today

Issued IP + control software

Four granted U.S. patents. The orchestration logic runs in browser platforms and a ROS 2 / Gazebo simulation package, with automated tests in continuous integration.

Research

Unattended autonomy

Production-grade unsupervised service. Multi-year research — the direction the portfolio protects, not a current capability.

Licensing · joint development · inventor support

Issued IP, working software, and a clear way to license it.

If your company applies films, coatings or lighting to vehicles, vessels or buildings — or builds the robots that do — there is a defined way to work with this portfolio. Technical and commercial questions are both welcome.

Product modules

Three modules. One platform.

Each module is backed by its own issued U.S. patent and is written to drive commercially available robotic platforms. A licence is granted per module and per field of use, not across the whole portfolio.

Everything described below is simulation and control software. No module has run on a physical robot, sensor or vehicle. The capabilities listed are what the software implements and what the claims recite — per-element implementation status is published at /patents.
Module 01
AutoWrap SURFACE
US 12,643,243 B1 · 10 claims
AI Platform for Automated Vehicle Surface Treatment Using Humanoid Robots
PPF · Vinyl Wrap · Ceramic Coating · Window Tint · Custom Graphics · Decals

AI-driven software that controls robotic platforms to apply paint protection film, vinyl wraps, ceramic coatings, window tint, custom graphics and decals on vehicles, watercraft and building glazing — with real-time 3D surface mapping, automated defect detection and self-correcting application. No proprietary hardware required.

Real-time 3D surface scan — stereo cameras and laser grid projectors map contours and panel edges before first tool contact
Automated multi-tool switching mid-process, including controllable heat-gun end effectors, heated spatula and knifeless trim
Integrated sensors detect bubbles, wrinkles and coating imperfections, and the system adjusts application parameters in response
Learns technique from expert human demonstration, mapped onto the robot’s joints by imitation learning and motion capture
Dedicated routine for vehicle mirror assemblies — micro-contour segmentation and adaptive stretch per zone
Runs on articulated arms, gantries, mobile platforms and humanoids; the specification names third-party humanoid platforms as compatible examples
License model
Per-Robot Monthly
Or per-vehicle royalty for fleet operators
Also available
OEM Integration
For robot manufacturers embedding the platform
Module 02
AutoWrap LIGHTS
US 12,649,240 B1 · 15 claims
AI-Led Platform for Automated Vehicle Treatment Using Humanoid Robots
Starlight Headliners · Underglow · Headlight / Taillight Service

Autonomous installation and servicing of vehicle lighting — starlight LED fibre-optic headliners, exterior underglow, and functional lighting. This patent contains the portfolio’s only ordered method claim: scan, plan, transport and prepare, install, route and verify wiring, then quality-control and correct.

3D interior scan generates the headliner geometry and fibre-optic insertion map before any material contact
End-effectors for fibre-optic manipulation, trim removal, wiring tools and lens-restoration polishing, selected and maintained autonomously
Sensors monitor placement, tension, LED alignment, beam pattern and wiring integrity
Beam aim measured against an acceptance box and corrected, then verified through the vehicle’s own diagnostic interface
Operates without continuous internet connectivity via on-device AI, and recovers autonomously from obstructions and minor tool malfunctions
On-site mobile deployment — fixed facilities and customer premises are both recited
License model
Per-Installation Royalty
Or monthly per-robot for high-volume operators
Also available
Luxury Auto Chains
Network-wide license for dealers + detailers
Module 03
AutoWrap SERVICE
US 12,511,598 B1 · 20 claims
Humanoid AI for Autonomous Vehicle Service Centers
Full Service Centre Automation · Diagnostics · Fleet Management

The autonomous vehicle service centre control layer. A centralised AI module orchestrates a fleet of humanoid robots across service centre functions — customer check-in, multi-modal diagnostics including EV and software-defined vehicles from model year 2005 onwards, repair, parts logistics, premises upkeep and customer communication.

Humanoid roles are recited by name: customer check-in, payment, diagnostics, physical repair, logistics, outdoor maintenance and security monitoring
Multi-modal sensor data and vehicle diagnostics feed explainable-AI (XAI) repair strategies, not just a fault code
One service plan dispatched across humanoids with dynamically assigned roles, and a digital twin of both vehicle and centre kept current
Robot operation constrained to user-defined geospatial boundaries; offline or malfunctioning units trigger task re-allocation
Humanoids drive vehicles into the bay, connect and disconnect EV charging, and schedule charge against grid demand and pricing
Multi-humanoid collaboration with joint manipulation, human-robot cooperation and AR-assisted human oversight
License model
Enterprise SaaS
Annual license per service centre location
Also available
Network License
Dealership groups, fleet operators, auto chains

Scope, pricing and terms are set per agreement. Nothing here is an offer, a licence, or legal advice. The demonstrations linked above are physics simulations, watermarked as such — not recordings of a real robot.