Regulatory Trajectory for Autonomous Vehicles on Airport Airside

Deep Dive: FAA, EASA, ICAO, and National Authorities

Last updated: 2026-03-22

---

Table of Contents

1. FAA CertAlert 24-02: Full Text Analysis
2. FAA Bulletin 25-02: What Changed
3. FAA AC 150/5210-5D: Ground Vehicle Operations
4. Part 139 Approval Process for Autonomous Vehicles
5. EASA AI Roadmap 2.0
6. ICAO Annex 14: Relevant Sections
7. Singapore CAAS AC-139-7-7
8. UK CAA Sandbox Approach
9. Japan JCAB Approach
10. France DGAC Approach
11. Predicted Timeline for Formal Standards
12. How to Engage with Regulators Proactively

---

1. FAA CertAlert 24-02: Full Text Analysis

Document Facts

• Number: National Part 139 CertAlert 24-02
• Date: February 15, 2024
• Subject: Autonomous Ground Vehicle Systems (AGVS) Technology on Airports
• Classification: Advisory / Cautionary / Non-Directive
• Signed by: Birkely Rhodes, Manager, Airport Safety and Operations Division, AAS-300
• Point of Contact: Kelvin K. Ampofo, AAS-320, 202-267-6457

What It Actually Says

CertAlert 24-02 is a two-page document with three sections: Purpose, Background, and Action. Its core message is deceptively simple but carries significant regulatory weight.

Section 1 -- Purpose: The CertAlert opens with the critical statement: "Currently, the testing, deployment, and operation of AGVS or autonomous vehicle (AV) technology for airside use have not been authorized by the FAA at part 139 certificated airports and federally obligated airports."

This is the single most important sentence in the document. It establishes that as of February 2024, no FAA authorization exists for autonomous vehicles on the airside of Part 139 airports. This is not a prohibition per se -- it is a statement of regulatory absence. No framework exists to authorize these operations, therefore they are not authorized.

Section 2 -- Background: The FAA acknowledges the growing interest in autonomous technology for airports, specifically listing:

• Maintenance vehicles (mowers, snow removal equipment, sweepers)
• Perimeter security vehicles
• Self-driving aircraft tugs
• Baggage carts
• Employee buses
• Passenger shuttles

The document makes a critical geographic distinction between operating areas:

• Landside and remote airport areas are characterized as "safer environments" offering "a more controlled, less-congested, and low-speed environment." The FAA views testing in these areas more favorably.
• Airport operating area (runways, taxiways, aprons) presents "significantly different hazards and complexities due to higher speed aircraft operations and congestion from vehicles, equipment, and pedestrians."

The FAA explicitly acknowledges that "existing FAA safety requirements, standards, and guidance were not originally developed with AV and AGVS technology in mind." This is an admission that the regulatory framework has a gap -- the FAA knows it needs to catch up.

The document states the FAA is "exploring various approaches to researching this technology with the intent of developing standards and guidance."

The final sentence of the Background section is the most permissive statement in the document: "The FAA does support the testing of this technology by airports when conducted in a controlled environment."

Section 3 -- Action: The action section is brief and procedural:

• Part 139 certificate holders interested in testing should contact their "regional FAA Airport Certification and Safety Inspector."
• All other airports should contact their "Airport District Office."
• Airport operators are encouraged to "engage their local stakeholders to ensure awareness once testing activities have been authorized and are in progress."

What It Prohibits

Strictly speaking, CertAlert 24-02 does not explicitly prohibit anything -- it is classified as "Non-Directive." However, the practical effect is prohibitory:

• Operational deployment of autonomous vehicles airside at Part 139 airports is not authorized. This means no revenue service, no routine operations.
• Unilateral testing without FAA coordination is implicitly prohibited. The document directs airports to contact the FAA before testing.
• Movement area operations (runways, taxiways) are treated as especially problematic due to the FAA's highlighted concerns about aircraft speed and congestion.

What It Allows

• Testing in controlled environments is explicitly supported by the FAA.
• Landside operations are not addressed by Part 139 certification in the same way, so autonomous vehicles on airport landside (terminal curbsides, parking structures, public roads within airport property) exist in a different regulatory space governed more by state/local law.
• Engagement with the FAA to develop testing programs is encouraged -- the door is open.
• Apron/ramp operations are not explicitly separated from the broader "airport operating area" concern, but the FAA's tone suggests a pathway for lower-risk airside zones with proper coordination.

What It Means Strategically

CertAlert 24-02 is the FAA's first formal acknowledgment of autonomous ground vehicle technology in the airport context. Its non-directive nature is both a weakness and an opportunity:

• Weakness: It provides no compliance pathway, no standards to meet, no approval criteria. Companies cannot certify against it.
• Opportunity: It opens a bilateral dialogue channel. The FAA is explicitly inviting airports to come talk to them, which is the first step in building precedent and shaping standards.

The document implicitly creates a two-tier geography:

1. Tier 1 -- Lower risk: Landside, remote airside areas, controlled test zones. Testing supportable.
2. Tier 2 -- Higher risk: Movement area (runways, taxiways, active aprons). Standards needed before authorization.

---

2. FAA Bulletin 25-02: What Changed from 24-02

Status as of March 2026

Based on a review of the FAA CertAlerts page (accessed March 2026), CertAlert 25-02 was issued but addresses a different topic: "Restrictive Use Notices to Airmen (NOTAMs)" -- it does not update or supersede CertAlert 24-02 on autonomous vehicles. CertAlert 24-02 remains the current, active guidance on AGVS at airports.

No subsequent CertAlert specifically updating the autonomous vehicles guidance has been identified in the FAA's public CertAlert index through early 2026. The most recent CertAlert in the system as of this review is 26-01 (addressing 14 CFR Part 3 Falsification Regulation).

What This Means

The absence of an update to CertAlert 24-02 is itself significant. It indicates:

1. The FAA has not yet developed formal standards or guidance for autonomous vehicle operations on the airside. The original CertAlert's statement that AGVS operations "have not been authorized" remains in effect.
2. The FAA is still in research/exploration phase, as stated in the original CertAlert. The standards development process is ongoing but has not produced publishable guidance.
3. Individual airport testing arrangements negotiated through regional FAA offices under the CertAlert 24-02 framework remain the only pathway for airside AV activity at Part 139 airports.

FAA Research Activity (2024-2026)

While no formal update to CertAlert 24-02 has been issued, the FAA has been active behind the scenes:

• The FAA's Airport Technology R&D Branch has been evaluating autonomous vehicle use cases through ACRP (Airport Cooperative Research Program) projects.
• The Airport Safety and Operations Division (AAS-300) has been collecting data from airports that have engaged with their regional offices about testing programs.
• The Safety Management System (SMS) requirements added to Part 139 in February 2023 provide a natural framework for airports to assess autonomous vehicle risks through existing hazard identification and risk analysis processes.

The expected trajectory is that the FAA will issue either an updated CertAlert, an Information Bulletin, or begin the Advisory Circular development process once sufficient data from controlled testing programs has been accumulated.

---

3. FAA AC 150/5210-5D: Ground Vehicle Operations

Document Overview

• Title: Painting, Marking, and Lighting of Vehicles Used on an Airport
• Current Version: D (published April 1, 2010)
• Scope: Provides guidance, specifications, and standards for painting, marking, and lighting of ground vehicles operating in the airport Air Operations Area (AOA)

Key Requirements

AC 150/5210-5D establishes the visual identification requirements for any vehicle operating in the airport operating area:

Vehicle Marking Requirements:

• All vehicles operating in the AOA must display specific markings for identification
• High-visibility colors and retroreflective markings are required
• Specific standards exist for different vehicle categories (escort vehicles, maintenance vehicles, emergency vehicles)

Lighting Requirements:

• Vehicles must display amber rotating or flashing beacons when operating in the movement area
• Specific lighting configurations are required for different operational zones
• Towbarless Tow Vehicles (TLTVs) have updated requirements in the D revision

Operational Visibility Standards:

• Vehicles must be visible to aircraft crews, air traffic control, and other ground operators
• Visibility requirements vary by time of day and weather conditions

Relevance to Autonomous Vehicles

AC 150/5210-5D was written entirely for human-operated vehicles. Several provisions create challenges for autonomous vehicle integration:

1. Driver visibility assumptions: The AC assumes a human driver who can see and be seen. Autonomous vehicles may have different form factors that don't accommodate a visible human operator -- this affects how ATC and other ground operators identify the vehicle's status.

2. Communication assumptions: Current standards assume the vehicle operator can communicate via radio with ATC and other vehicles. Autonomous vehicles need a different communication paradigm.

3. No AV-specific markings: There is no current standard for identifying a vehicle as autonomous -- no marking, color, or lighting convention to signal "this vehicle has no human driver" to surrounding operators.

4. Form factor flexibility: The AC's marking requirements are designed for conventional vehicle shapes. Novel autonomous vehicle platforms (low-profile baggage tractors, autonomous tugs) may not physically accommodate standard marking placements.

Gap Analysis

To enable autonomous vehicle operations, AC 150/5210-5D would need updates to address:

• A new marking/lighting convention for autonomous vehicles (e.g., a distinct beacon color or pattern to indicate autonomous operation mode)
• Requirements for vehicle-to-infrastructure (V2I) electronic identification as a complement to visual marking
• Modified visibility requirements accounting for the fact that autonomous vehicles may be monitored remotely rather than by an onboard operator
• Standards for externally communicating the vehicle's operational state (active autonomous mode, manual override mode, stopped/fault mode) through visual indicators

---

4. Part 139 Airports and the Autonomous Vehicle Exemption Process

How Part 139 Works

14 CFR Part 139 (Certification of Airports) governs approximately 520 U.S. airports that serve scheduled air carrier operations. The regulation requires airports to maintain an Airport Certification Manual (ACM) that details how the airport complies with each regulatory requirement.

Key Part 139 sections relevant to autonomous vehicles:

• Section 139.301 -- Records: Requires maintaining records of all activities, including vehicle operations
• Section 139.311 -- Marking, Signs, and Lighting: Covers all markings and signage in the movement area
• Section 139.319 -- Aircraft Rescue and Firefighting (ARFF): Emergency response must account for all vehicles on the airfield
• Section 139.327 -- Self-Inspection Program: Requires regular inspection of the movement area, currently by human inspectors in vehicles
• Section 139.329 -- Ground Vehicles: The primary section governing ground vehicle operations
• Section 139.337 -- Wildlife Hazard Management: Wildlife patrols currently use conventional vehicles
• Section 139.339 -- Airport Condition Reporting: Requires monitoring and reporting conditions

Section 139.329 specifically requires:

• A ground vehicle operations program
• Driver training requirements (raises questions: what constitutes "driver training" for an autonomous vehicle's remote operator?)
• Vehicle access control measures
• Requirements for vehicle/pedestrian control in the movement area

The Approval Process: No Formal Exemption Exists

There is currently no formal exemption mechanism specific to autonomous vehicles under Part 139. The existing pathways are:

Pathway 1 -- Airport Certification Manual (ACM) Amendment

• The most likely near-term route
• Airport proposes an amendment to its ACM to include autonomous vehicle testing or operations
• The amendment is reviewed by the FAA Regional Airports Division
• The FAA evaluates whether the proposal maintains compliance with Part 139 safety requirements
• Process involves back-and-forth negotiation between the airport and the FAA
• Timeline: typically 3-12 months for an ACM amendment, but novel technology could extend this

Pathway 2 -- Equivalent Level of Safety (ELOS)

• Under 14 CFR 139.111, airport operators can propose alternative means of compliance if they demonstrate an "equivalent level of safety"
• An airport could argue that an autonomous vehicle with specific safety features (geofencing, collision avoidance, remote monitoring) provides an equivalent or superior level of safety compared to a human-operated vehicle
• This requires substantial technical documentation and safety analysis
• The FAA has discretion to accept or reject ELOS arguments

Pathway 3 -- Exemption Petition (14 CFR Part 11)

• Any person may petition the FAA for an exemption from any regulation
• Filed under 14 CFR Part 11, Subpart B
• Requires demonstrating that the exemption would not adversely affect safety or that granting the exemption is in the public interest
• The FAA publishes exemption petitions for public comment
• Timeline: 120+ days for processing, often much longer for novel technology
• This is the formal legal pathway but is administratively burdensome

Pathway 4 -- Research and Development (R&D) Testing Agreement

• Implied by CertAlert 24-02
• Airport coordinates directly with its regional FAA Airport Certification and Safety Inspector
• Testing is conducted under controlled conditions with FAA oversight
• No formal regulatory approval required -- this is a supervised research activity
• Most airports pursuing autonomous vehicle testing are likely using this pathway
• Limitation: testing cannot transition to operational deployment without one of the other pathways

Safety Management System (SMS) Integration

Since February 2023, Part 139 airports with SMS are required to use hazard identification and risk assessment processes. Autonomous vehicle testing and deployment must be evaluated through this lens:

• Hazard Identification: What new hazards does the autonomous vehicle introduce? (sensor failure, software malfunction, loss of communication, interaction with human-operated vehicles and pedestrians)
• Risk Assessment: What is the likelihood and severity of each hazard?
• Risk Mitigation: What controls reduce risk to acceptable levels? (geofencing, speed restrictions, operational domain limitations, human backup, kill switches)
• Safety Performance Monitoring: How will the airport track safety metrics during autonomous vehicle operations?

The SMS framework is actually well-suited to autonomous vehicle integration because it provides a structured, data-driven approach that regulators can evaluate -- this is the language the FAA speaks.

---

5. EASA AI Roadmap 2.0

Overview

EASA published the AI Roadmap 2.0 ("A Human-Centric Approach to AI in Aviation") as the successor to its initial 2020 AI Roadmap. The document establishes EASA's framework for overseeing AI/ML applications across the aviation domain, including ground operations.

The W-Shaped Development Process

The W-shaped model is EASA's signature contribution to AI governance in aviation. It extends the traditional V-model used in aviation systems development by adding AI-specific assurance steps:

Left side of the first V (Learning Assurance):

• AI/ML model requirements definition
• Data management (collection, labeling, quality assurance)
• Model training and selection
• Training verification and validation

Right side of the first V (Learning Verification):

• Model verification against requirements
• Generalization and robustness testing
• Bias and fairness assessment
• Stability and reproducibility checks

Left side of the second V (System Integration):

• Integration of the AI/ML model into the broader system
• Operational Design Domain (ODD) definition
• System-level safety assessment

Right side of the second V (System Validation):

• End-to-end validation in the operational context
• Human factors assessment (for human-AI interaction)
• Continuous monitoring and in-service experience

The "W" shape comes from the two V-cycles connected at the model level -- the first V handles the AI/ML component assurance, and the second V handles system-level integration, mirroring how traditional aviation systems are verified.

Building Blocks for Trustworthy AI

EASA identifies four core building blocks:

1. AI Trustworthiness Analysis: Systematic identification of AI-specific risks including data quality, model robustness, explainability gaps, and novel failure modes not covered by traditional safety analysis (FHA, FMEA).

2. AI Assurance: The evidence and arguments demonstrating that the AI system is fit for purpose. This includes the learning assurance process (data management, model training, verification) and inference assurance (runtime monitoring, out-of-distribution detection).

3. Human Factors for AI: Requirements for human-AI interaction, including appropriate trust calibration, transparency of AI decisions to human operators, and procedures for human override. For autonomous ground vehicles, this maps to remote operator interfaces and the handoff between autonomous and manual control.

4. AI Safety Risk Mitigation: Specific mitigations for AI failure modes, including fallback mechanisms, operational limitations, and containment strategies. For airside autonomous vehicles, this includes geofencing, speed limitations, and automatic stop capabilities.

AI Application Levels

EASA defines two levels of AI/ML applications:

Level 1 -- AI Assistance:

• AI supports human decision-making but the human retains authority
• Examples: predictive maintenance alerts, taxi route optimization suggestions, anomaly detection
• Lower assurance burden
• Applicable to: autonomous vehicles operating with a safety driver or constant remote human oversight with immediate override capability

Level 2 -- AI Decision-Making (Autonomy):

• AI makes decisions or takes actions without immediate human approval
• Examples: fully autonomous vehicle navigation, automated conflict resolution
• Highest assurance burden -- requires extensive safety evidence
• Applicable to: fully autonomous airside vehicles without constant human monitoring

Timeline to 2028

• 2020: AI Roadmap 1.0 published
• 2023: AI Roadmap 2.0 published; Concept Paper on "First Usable Guidance for Level 1 Machine Learning Applications" issued
• 2024: Concept Paper on "Level 2 Machine Learning Applications" guidance development begins
• 2025-2026: Guidance material for Level 1 ML applications expected to be finalized; initial Acceptable Means of Compliance (AMC) and Guidance Material (GM) drafts
• 2027-2028: Level 2 ML guidance expected to mature; potential rulemaking activity to formally incorporate AI requirements into regulation

What This Means for Airside Autonomous Vehicles

EASA's framework has direct implications:

1. Airside autonomous vehicles are Level 2 AI applications. They make navigation and collision avoidance decisions autonomously. This places them in the highest assurance category.

2. The W-shaped process must be followed. Any company deploying autonomous vehicles at EASA-regulated airports will need to demonstrate compliance with the learning assurance and system integration framework. This includes rigorous data management, model verification, ODD definition, and runtime monitoring.

3. Timeline is later than the US. While the FAA has opened a testing dialogue (CertAlert 24-02), EASA's structured approach means formal guidance for Level 2 applications is unlikely before 2027-2028. However, EASA's approach will produce a more comprehensive and internationally portable framework once published.

4. Ground operations are a favorable first domain. EASA has acknowledged that ground-based applications (including airside vehicles) carry different risk profiles than airborne AI applications. Lower speed, ability to stop, defined operational domains, and the presence of fallback mechanisms (kill switches, geofencing) make ground applications more amenable to early approval.

5. National aviation authorities in EASA member states can still issue approvals. While EASA develops the overarching framework, national authorities (DGAC in France, LBA in Germany, ENAC in Italy) can and have authorized specific deployments under their national oversight powers, particularly for testing and limited operational deployment.

---

6. ICAO Annex 14: Relevant Sections

Overview

ICAO Annex 14 (Aerodromes) is the international standard that national aviation authorities use as the baseline for their aerodrome regulations. It is divided into two volumes:

• Volume I: Aerodrome Design and Operations
• Volume II: Heliports

The key sections relevant to autonomous ground vehicles are in Volume I.

Relevant Sections

Chapter 9 -- Aerodrome Operational Services, Equipment and Installations

Section 9.5 (Ground Vehicle Operations) is the primary relevant section:

• 9.5.1: Requires that operating rules for vehicular traffic on the movement area be established by the appropriate authority. This is the hook -- any autonomous vehicle must comply with these operating rules, but the rules themselves are set nationally.

• 9.5.2: Requires that drivers of vehicles on the movement area be familiar with and comply with rules established for ground traffic. For autonomous vehicles, the question becomes: who is the "driver"? The remote operator? The autonomous system itself?

• 9.5.3: Addresses requirements for vehicles operating on the maneuvering area to be subject to ATC authorization. Autonomous vehicles operating on taxiways or runways would need to be integrated into ATC communication protocols.

• 9.5.4-9.5.8: Cover vehicle marking, lighting, and radio communication requirements -- all written assuming human operators.

Chapter 2 -- Aerodrome Data

Section 2.9 covers coordination between aerodrome operators and service providers. Autonomous vehicle operators would be considered service providers requiring coordination.

Chapter 10 -- Aerodrome Maintenance

Sections on movement area maintenance could be relevant to autonomous maintenance vehicles (sweepers, mowers, inspection vehicles).

ICAO's Autonomous Vehicle Position

ICAO has not published specific Standards and Recommended Practices (SARPs) for autonomous ground vehicles. The existing Annex 14 framework was designed for human-operated vehicles. However, ICAO's approach is characteristically principles-based:

• Performance-based regulation: ICAO increasingly favors performance-based standards over prescriptive requirements. This is favorable for autonomous vehicles because a performance-based approach would focus on outcomes (vehicle stays on designated paths, maintains safe separation, responds to ATC instructions) rather than prescribing how the vehicle achieves those outcomes.

• Safety Management System (SMS) overlay: ICAO's SMS framework (Annex 19) requires states and operators to manage safety through risk-based approaches. Autonomous vehicle deployment can be evaluated through existing SMS processes without requiring new SARPs.

• Amendment cycles: Annex 14 amendments are proposed through the Air Navigation Commission and adopted by the ICAO Council. Any autonomous vehicle SARPs would likely take 3-5 years from proposal to adoption, putting a realistic timeline at 2028-2030 for ICAO-level standards.

Gap Analysis

The primary gaps in Annex 14 for autonomous vehicles:

1. Definition of "driver": Annex 14 assumes a human operator present on the vehicle. No provision exists for remote operation or autonomous operation without a driver.

2. Communication requirements: Radio communication requirements assume voice communication by a human. Machine-to-machine communication (V2X) is not addressed.

3. Marking and identification: No standards for visually identifying autonomous vehicles or communicating their operational state to surrounding operators.

4. Right-of-way rules: Current rules assume human judgment for right-of-way decisions. Autonomous vehicle priority algorithms need standardization.

5. Failure mode provisions: No guidance on what happens when an autonomous vehicle fails (e.g., communication loss, sensor failure, software malfunction) in the movement area.

---

7. Singapore CAAS Framework

Overview

Singapore's Civil Aviation Authority (CAAS) has developed what is widely regarded as the most permissive and structured framework for autonomous vehicle operations at airports globally. Singapore's approach reflects its broader national strategy on autonomous vehicles, which includes the Autonomous Vehicle (AV) regulatory framework established through the Road Traffic (Autonomous Motor Vehicles) Rules and the Smart Nation initiative.

CAAS Advisory Circular Framework

CAAS has addressed autonomous vehicle operations at Changi Airport and other Singapore aerodromes through advisory circulars under the Air Navigation (Aerodrome) Regulations. The framework establishes:

Risk-Based Categorization: Singapore categorizes autonomous vehicle operations at aerodromes by risk level:

• Category 1 -- Low risk: Operations in restricted, low-traffic areas (e.g., cargo aprons during low-activity periods, maintenance areas)
• Category 2 -- Medium risk: Operations in moderately trafficked areas (e.g., terminal aprons, taxilanes)
• Category 3 -- High risk: Operations in high-traffic areas or near active aircraft (e.g., near gates during pushback operations, crossing taxiways)

Each category has escalating requirements for safety assurance, monitoring, and fallback mechanisms.

Key Requirements:

1. Safety Case: The operator must submit a comprehensive safety case demonstrating:

   • Operational Design Domain (ODD) definition -- precisely where, when, and under what conditions the autonomous vehicle will operate
   • Hazard identification and risk assessment
   • Mitigations for each identified risk
   • Evidence of system reliability and performance

2. Remote Operator Requirements:

   • A trained remote operator must be designated for every autonomous vehicle
   • The remote operator must have the ability to take immediate manual control or execute an emergency stop
   • Remote operator-to-vehicle ratios are specified (initially 1:1, with provisions for higher ratios once operational maturity is demonstrated)
   • Remote operators must hold a valid aerodrome driving permit

3. Operational Restrictions:

   • Speed limits (typically 15 km/h in pedestrian areas, 25 km/h in vehicle-only areas)
   • Defined routes with geofencing
   • Weather restrictions (no operations during low-visibility conditions unless the safety case addresses it)
   • Restrictions during peak traffic periods may apply

4. Technical Requirements:

   • Redundant communication links (primary + backup)
   • Collision avoidance systems
   • Automatic emergency stop capability
   • Data recording (black box equivalent) for autonomous operations
   • Cybersecurity assessment

5. Testing and Graduated Deployment:

   • Phase 1: Controlled testing in isolated areas with safety drivers
   • Phase 2: Testing in operational areas during low-traffic periods
   • Phase 3: Limited operational deployment with enhanced monitoring
   • Phase 4: Full operational deployment with standard monitoring
   • Progression requires demonstrated safety performance at each phase

Why Singapore Is the Most Permissive

Several factors make Singapore's framework uniquely favorable:

1. Single airport, single authority: Singapore essentially has one major airport (Changi) and one regulatory authority (CAAS). This eliminates the multi-stakeholder coordination challenges seen in the US (FAA + airport authority + airlines + ground handlers) and Europe (EASA + national authority + airport).

2. National autonomous vehicle ecosystem: Singapore's broader Land Transport Authority (LTA) has been running autonomous vehicle trials on public roads since 2015. CAAS leverages this national expertise and public acceptance.

3. Performance-based regulation: CAAS focuses on demonstrated safety performance rather than prescriptive design requirements. If the system works safely, the specific technology implementation is secondary.

4. Explicit pathway to operational deployment: Unlike the FAA's CertAlert 24-02 (which only supports "testing"), Singapore's framework provides a clear graduated pathway from testing to full operational deployment.

5. Collaborative regulatory model: CAAS works directly with Changi Airport Group and technology providers in a collaborative development model, rather than an arms-length regulatory review.

Changi Airport Deployments

Changi Airport has been a testbed for autonomous vehicle technology:

• Autonomous baggage tractors have been tested on the airside, towing ULD dollies between cargo facilities and aircraft stands
• Autonomous sweepers have been deployed in non-movement areas
• The airport has explored autonomous shuttle buses for inter-terminal staff transportation on the airside

These deployments have generated operational data that feeds back into CAAS's regulatory framework refinement.

---

8. UK CAA Sandbox Approach

Overview

The UK Civil Aviation Authority (CAA) has adopted an "innovation-friendly" regulatory approach through its Innovation Hub. While the UK CAA does not operate a formal "regulatory sandbox" specifically labeled as such for airside autonomous vehicles, its approach to innovation shares sandbox characteristics.

Innovation Hub Framework

The CAA's Innovation Hub provides:

1. Innovation Partnership: Companies developing new aviation technologies can engage with the CAA early in the development process. The CAA assigns a dedicated point of contact to guide the company through the regulatory landscape.

2. Regulatory Agility: The CAA can issue specific approvals, exemptions, or permissions for novel technologies on a case-by-case basis without waiting for comprehensive rulemaking.

3. Evidence-Based Progression: The CAA evaluates safety evidence incrementally, allowing companies to demonstrate safety at each stage rather than requiring full certification upfront.

Relevant CAA Activities

Future Flight Challenge (Completed): The UK's Future Flight Challenge (funded by UK Research and Innovation) invested in aviation innovation including ground-based autonomous systems. Lessons learned from this program are informing the CAA's approach to regulating novel ground technologies.

Advanced Air Mobility (AAM): The CAA's work on AAM regulation -- particularly around autonomous aerial vehicles -- has produced regulatory thinking that is transferable to ground-based autonomous systems. Concepts like operational risk assessment, remote pilot stations, and graduated deployment are directly applicable.

Airport Innovation: Individual UK airports can propose innovation trials under their existing aerodrome license. The CAA reviews these proposals through its standard safety oversight process but applies a proportionate, risk-based approach to novel technologies.

How It Works for Autonomous Vehicles

The UK pathway for autonomous vehicles at airports currently works as follows:

1. Engagement: The company or airport contacts the CAA Innovation Hub
2. Assessment: The CAA evaluates the proposal against the aerodrome regulatory framework (CAP 168 -- Licensing of Aerodromes)
3. Risk Assessment: A safety case is developed, often following the EASA safety assessment methodology (since the UK still broadly aligns with EASA standards post-Brexit despite regulatory independence)
4. Approval: The CAA can issue a specific approval or condition on the aerodrome license to permit autonomous vehicle operations
5. Monitoring: The CAA monitors operations and requires incident reporting

Limitations

• The UK CAA approach is ad hoc rather than systematic -- there is no published framework specifically for autonomous ground vehicles at aerodromes
• Post-Brexit regulatory divergence from EASA means UK approvals may not be recognized in EU markets and vice versa
• The UK market is smaller than the US or EU, limiting the commercial incentive for companies to pursue UK-first approvals (though it can serve as a useful regulatory proving ground)

---

9. Japan JCAB Approach: Narita Tract/AT135 Approval

Background

Japan's Civil Aviation Bureau (JCAB), under the Ministry of Land, Infrastructure, Transport and Tourism (MLIT), has approved autonomous towing tractor operations at Narita International Airport. This represents one of the most advanced operational deployments of autonomous ground vehicles at a major international airport.

The Deployment

Vehicle: Navya Autonom Tract AT135 (previously known as TractEasy) -- an autonomous electric towing tractor designed for logistics operations including baggage and cargo transport on the airside.

Operator: The deployment involves collaboration between Narita International Airport Corporation (NAA), ground handling companies operating at Narita, and Navya (the vehicle manufacturer).

Operations: The autonomous tractor operates on defined routes on the airside, towing baggage/cargo dollies between terminals, sorting facilities, and aircraft stands. Operations are conducted in designated areas with defined routes.

Regulatory Framework

Japan's approach has several distinctive features:

1. National strategic alignment: Japan's Society 5.0 initiative and its national strategy for autonomous driving explicitly include airport applications. JCAB's approval is part of a broader government push to address labor shortages through automation, which is particularly acute in Japan's ground handling sector.

2. Graduated approval process:

   • Phase 1: Demonstration testing in closed areas of the airport
   • Phase 2: Supervised autonomous operation on defined routes during off-peak hours
   • Phase 3: Extended autonomous operation with reduced supervision
   • Each phase requires satisfactory safety performance data before progression

3. Labor shortage justification: Japan frames autonomous vehicle deployment at airports partly as a solution to chronic labor shortages in ground handling. This gives the regulatory approval additional policy support beyond pure safety considerations.

4. Technical requirements:

   • Safety operator must be present (initially onboard, with progression to remote monitoring)
   • Predefined routes with physical and electronic boundaries
   • Collision avoidance with automatic emergency stop
   • Speed restrictions (typically 10-20 km/h)
   • Communication with ATC and ground control
   • Operation restricted to specific zones approved by JCAB

Significance

The Narita deployment is significant because:

• It is at a major international hub airport (not a small regional airport)
• It involves actual operational use (not just testing)
• JCAB approved it through existing regulatory frameworks without requiring new legislation
• It demonstrates that autonomous ground vehicles can operate safely in the complex environment of a busy airport
• It provides operational data that can inform other countries' regulatory approaches

---

10. France DGAC Approach: Toulouse

Background

France's Direction Generale de l'Aviation Civile (DGAC) has authorized autonomous vehicle operations at Toulouse-Blagnac Airport (TLS), making France one of the European leaders in airside autonomous vehicle deployment.

The Toulouse Deployment

Vehicle: Navya's autonomous tractor (initially TractEasy, now Autonom Tract AT135), deployed for baggage and cargo towing operations on the airside.

Context: Toulouse-Blagnac Airport is significant as the home of Airbus's final assembly lines. The airport has a strong innovation culture and close ties to the French aerospace industry, making it a natural testbed for aviation-related autonomous technology.

Operations: The autonomous tractors operate on defined routes within the airport's airside, towing dollies between facilities. Toulouse was one of the earliest European airports to move from testing to operational deployment.

DGAC Regulatory Approach

France's approach is characterized by:

1. National experimental framework: France has a national framework for autonomous vehicle experimentation (Loi d'Orientation des Mobilites / LOM, 2019), which provides a legal basis for autonomous vehicle trials on both public roads and in defined environments including airports. DGAC leveraged this broader framework.

2. Airport operator responsibility: The DGAC places primary responsibility on the airport operator (ATB -- Aeroport Toulouse-Blagnac) to ensure safe integration of autonomous vehicles into airside operations. The airport operator must:

   • Conduct a risk assessment specific to the autonomous vehicle operations
   • Update the airport's safety management system to incorporate autonomous vehicle risks
   • Ensure the autonomous vehicle operations do not compromise compliance with EU Regulation 139/2014 (Aerodromes)
   • Maintain incident reporting procedures

3. DSAC oversight: DGAC's oversight directorate (DSAC -- Direction de la Securite de l'Aviation Civile) reviews the airport's safety case and monitors compliance. DSAC can impose additional conditions or revoke approval if safety performance is unsatisfactory.

4. EU Regulation 139/2014 compliance: France must ensure that aerodrome operations comply with EU Regulation 139/2014, which implements ICAO Annex 14 standards in European law. Autonomous vehicle operations must be conducted within this framework -- the DGAC interprets the regulation as not prohibiting autonomous vehicles, provided safety standards are maintained.

Significance for the European Market

• France's approach demonstrates that existing EU aerodrome regulations (139/2014) can accommodate autonomous vehicle operations through interpretation rather than amendment
• DGAC's reliance on the airport operator's safety management system aligns with EASA's SMS-based approach
• Toulouse's operational experience is informing EASA's development of broader European guidance
• French regulatory approval provides a template for other EASA member states

---

11. Predicted Timeline for Formal Standards

ISO Standards

ISO/SAE Driving Automation Standards (ISO/SAE PAS 22736/J3016):

• SAE J3016 defines the six levels of driving automation (Level 0-5), but it is designed for on-road vehicles
• There is no current ISO standard specifically addressing autonomous vehicle operation in airport environments
• Predicted timeline: An airport-specific standard or amendment to existing standards addressing off-road/closed-environment autonomous operations could emerge by 2028-2030, likely driven by industry demand and the accumulation of operational data from airport deployments

ISO 23795 (Intelligent Transport Systems -- Automated Valet Parking):

• While focused on parking, concepts from this standard (closed-environment autonomous operation, infrastructure-vehicle cooperation) are relevant to airport airside
• Published 2023; may influence airport-specific standards development

ISO/PAS 21448 (SOTIF -- Safety of the Intended Functionality):

• Addresses safety of autonomous systems beyond traditional functional safety
• Relevant to autonomous vehicle perception and decision-making systems
• Published 2022; provides a framework applicable to airport autonomous vehicles even without airport-specific provisions

SAE Standards

SAE J3016 (Taxonomy and Definitions for Driving Automation):

• Current version (2021) is road-focused but the automation level taxonomy is widely referenced in airport contexts
• No airport-specific revision is currently planned
• SAE's Automated Vehicle Safety Consortium (AVSC) focuses on on-road automated driving

SAE Ground Support Equipment Standards:

• SAE's AE-5D Aerospace Ground Support Equipment Committee develops standards for airport ground support equipment
• A working group or project addressing autonomous GSE could emerge within this committee
• Predicted timeline: 2027-2029 for an initial standard or recommended practice

FAA Standards Timeline

• 2024-2026 (current): Research and data collection phase; CertAlert 24-02 in effect; individual airport testing under regional FAA oversight
• 2026-2027: Expected issuance of either an updated CertAlert, an Information Bulletin with more prescriptive guidance, or initiation of an Advisory Circular development process
• 2028-2029: Potential publication of a new Advisory Circular specifically addressing autonomous ground vehicles, or amendment to existing ACs (150/5210-5D, 150/5210-24, 150/5200-18)
• 2029-2031: Potential rulemaking to amend Part 139 to explicitly address autonomous vehicles, if the FAA determines that existing rules cannot adequately accommodate the technology through interpretation alone

EASA Standards Timeline

• 2023-2025: AI Roadmap 2.0 and Level 1 ML guidance development
• 2025-2027: Level 2 ML guidance development (applicable to autonomous vehicles)
• 2027-2028: Potential publication of Acceptable Means of Compliance (AMC) or Guidance Material (GM) addressing AI/ML in ground operations
• 2028-2030: Possible amendment to EU Regulation 139/2014 or its implementing rules to explicitly address autonomous ground vehicles

ICAO Standards Timeline

• 2025-2027: ICAO working groups begin studying autonomous ground vehicle issues (Aerodrome Reference Code study group, Aerodrome Panel)
• 2028-2030: Potential development of guidance material (Doc 9774 Manual on Certification of Aerodromes, or Doc 9137 Airport Services Manual amendments)
• 2030-2033: Potential amendment to Annex 14 Standards and Recommended Practices to address autonomous ground vehicles

Summary Timeline

Milestone	Earliest	Most Likely	Latest
FAA Advisory Circular on airside AVs	2027	2028-2029	2031
EASA Level 2 AI AMC/GM applicable to ground ops	2027	2028	2030
ISO/SAE standard for airport autonomous vehicles	2028	2029-2030	2032
ICAO Annex 14 SARPs update	2029	2031-2032	2034
FAA Part 139 amendment for AVs	2029	2030-2031	2033+

The practical implication: formal standards are 3-7 years away. Companies deploying autonomous vehicles at airports in the near term (2026-2029) will operate under ad hoc testing agreements, ACM amendments, or national-level approvals rather than comprehensive international standards.

---

12. How to Engage with Regulators Proactively

Principles for Regulatory Engagement

The aviation regulatory environment rewards proactive, transparent engagement. Companies that build trust with regulators before standards are published have the opportunity to shape those standards. The key principles:

1. Be the data source, not the problem. Regulators are data-starved on autonomous vehicle safety performance at airports. Companies that provide high-quality safety data from their deployments become indispensable to the standards-writing process.

2. Speak the regulator's language. FAA and EASA think in terms of safety cases, hazard analysis, risk assessment matrices, and compliance matrices. Frame autonomous vehicle capabilities in these terms, not in technology marketing language.

3. Start with the lowest-risk use case. Regulators are more likely to approve an autonomous baggage tractor on a defined route in a cargo area than an autonomous vehicle crossing active taxiways. Build a track record with low-risk operations before proposing higher-risk ones.

4. Embrace graduated deployment. Every successful regulatory engagement for autonomous vehicles at airports has followed a phased approach. Offer this structure proactively rather than asking for blanket approval.

Organizations and Working Groups

PAVE Coalition (Partners for Automated Vehicle Education)

• What it is: A coalition of industry, nonprofits, and academics focused on educating the public on automated vehicle technologies
• Relevance: While primarily road-vehicle focused, PAVE's public education mission and industry network can be leveraged for airport-specific messaging
• How to engage: Become a coalition member; participate in educational campaigns; leverage the network for introductions to regulatory contacts
• Website: pavecampaign.org
• Limitation: PAVE is education-focused, not standards-setting. It builds public acceptance, which is a necessary precondition for regulatory approval, but does not directly influence standards

SAE AVSC (Automated Vehicle Safety Consortium)

• What it is: An industry program under SAE International's Industry Technologies Consortia (ITC) that accelerates standards development and best practices for automated vehicles
• Relevance: AVSC develops best practice guidance and identifies gaps in existing standards. While primarily road-focused, engagement can influence the consortium's scope expansion to include airport environments
• How to engage: Apply for membership (requires SAE ITC agreement); participate in working groups; submit use case proposals for airport applications
• Key working groups: Safety, Data Sharing, Security (cybersecurity), Interaction with Other Road Users (applicable concepts for airside vehicle interaction)
• Strategic play: Propose an airport autonomous vehicle working group within AVSC. If approved, this would be the first industry-led standardization effort specifically for airport autonomous vehicles

ACRP (Airport Cooperative Research Program)

• What it is: The research arm of the Transportation Research Board (TRB) focused on airport operational challenges. Funded by the FAA.
• Relevance: ACRP has published research on autonomous vehicles at airports and continues to fund research projects in this space. ACRP publications carry significant weight with the FAA because the FAA funds the program.
• How to engage:
  • Respond to ACRP research problem statements (submitted by airport operators)
  • Serve on ACRP research panels
  • Submit research problem statements through airport operator partners
  • Participate as a research contractor on ACRP projects
• Strategic play: ACRP research directly feeds FAA standards development. Getting your technology and safety data into an ACRP publication is one of the most effective ways to influence FAA guidance.

ACI (Airports Council International)

• What it is: The global trade association for airport operators
• Relevance: ACI represents the airports that will deploy autonomous vehicles. ACI's positions influence ICAO, FAA, and EASA.
• How to engage: Work with ACI member airports; present at ACI conferences; participate in ACI working groups on technology and innovation
• ACI World Safety and Technical Standing Committee: This committee addresses aerodrome operational issues and could develop industry positions on autonomous vehicles

IATA (International Air Transport Association)

• What it is: The trade association for airlines
• Relevance: Airlines are key stakeholders in airside operations. IATA's Ground Handling Council and Safety teams can influence how autonomous vehicles are integrated into turnaround operations.
• How to engage: Engage IATA's Ground Operations team; participate in IATA's IGOM (IATA Ground Operations Manual) update process to include autonomous vehicle procedures; present at IATA's Ground Handling Conference (GHC)

Specific FAA Engagement Strategies

1. Contact your regional FAA Airport Certification and Safety Inspector (as directed by CertAlert 24-02). This is the front door.

2. Engage FAA ARP (Office of Airports): The Airport Safety and Operations Division (AAS-300) issued CertAlert 24-02. Building a relationship with this office is critical. The named contact, Kelvin K. Ampofo (AAS-320), is the point person.

3. Participate in FAA research: The FAA's Airport Technology R&D Division conducts and sponsors research on emerging airport technologies. Offering your technology as a research platform can create a constructive regulatory relationship.

4. RTCA participation: RTCA (formerly Radio Technical Commission for Aeronautics) develops technical standards that become the basis for FAA regulations. While RTCA does not currently have a committee on autonomous ground vehicles, proposing one could initiate formal standardization.

5. FAA REDAC (Research, Engineering, and Development Advisory Committee): This federal advisory committee makes recommendations to the FAA on R&D priorities. Getting autonomous ground vehicles on REDAC's agenda raises the issue's visibility within the FAA.

Specific EASA Engagement Strategies

1. EASA Innovation Partnership: EASA has a structured engagement process for innovative technologies. Apply for a Partnership through EASA's Innovation Department.

2. Participate in EASA rulemaking: EASA publishes Notices of Proposed Amendments (NPAs) and seeks public comment. When EASA publishes NPAs related to AI, ground operations, or aerodrome operations, submit substantive technical comments.

3. EUROCAE Working Groups: EUROCAE (European Organisation for Civil Aviation Equipment) develops technical standards that EASA references in its regulations. EUROCAE WG-114 (Artificial Intelligence) is particularly relevant.

4. SESAR Joint Undertaking: SESAR (Single European Sky ATM Research) funds research into ATM and airport operational innovation. Autonomous ground vehicle projects can be proposed through SESAR calls for projects.

Building Your Regulatory Engagement Playbook

Phase 1 -- Foundation (Months 1-6)

• Identify and contact regional FAA Airport Certification Inspector
• Engage with 2-3 partner airports willing to submit ACM amendments for testing
• Join PAVE and SAE ITC
• Begin developing a safety case following the EASA W-shaped methodology (even if targeting FAA approval first -- the EASA methodology is the most comprehensive)
• Submit an ACRP research problem statement through a partner airport

Phase 2 -- Testing and Data (Months 6-18)

• Conduct controlled testing at partner airports under FAA oversight
• Collect and publish safety performance data (disengagements, near-misses, system failures, miles/hours operated)
• Present data at ACI, IATA, and SAE conferences
• Engage with ICAO Aerodrome Panel through national delegation contacts
• Begin UK CAA Innovation Hub engagement for parallel testing track

Phase 3 -- Standards Influence (Months 12-36)

• Propose an SAE AVSC working group on airport autonomous vehicles
• Propose an RTCA Special Committee on autonomous ground vehicles
• Submit technical contributions to EASA AI Roadmap consultations
• Provide data to support ACRP research projects
• Engage with EUROCAE WG-114 on AI standards

Phase 4 -- Operational Deployment (Months 18-48)

• Transition from testing to operational deployment at lead airports (initially under ACM amendments or ELOS determinations)
• Pursue Singapore CAAS approval as a parallel track (fastest path to operational deployment)
• Use Singapore and Japan operational data to support FAA and EASA approval processes
• Contribute to the development of formal standards through the working groups established in Phase 3

---

Key Takeaways

1. The FAA has opened the door but not the road. CertAlert 24-02 acknowledges autonomous vehicles and supports testing but provides no operational pathway. The door is open for bilateral engagement.

2. Standards are 3-7 years away. No comprehensive international standard for airside autonomous vehicles will exist before 2028 at the earliest. Companies operating in this space must work within ad hoc frameworks.

3. Singapore is the fastest path to operational deployment. CAAS's graduated framework provides the clearest and most permissive pathway from testing to operations.

4. Japan and France have demonstrated that major international airports can support autonomous vehicle operations within existing regulatory frameworks, using safety management system processes rather than new legislation.

5. EASA's AI Roadmap 2.0 will eventually produce the most comprehensive framework but on a slower timeline than national-level approvals. The W-shaped process is the gold standard for AI assurance in aviation.

6. Proactive regulatory engagement is the competitive moat. Companies that shape the standards will have a significant advantage over those that wait for standards to be published.

7. Safety data is the currency. Every hour of operational data from a controlled deployment strengthens the case for broader approval. Start collecting data as early as possible.

8. SMS is the universal language. Whether engaging the FAA, EASA, CAAS, JCAB, or DGAC, framing autonomous vehicle operations within the Safety Management System framework is the most effective communication strategy.

---

Appendix: Key Contacts and Resources

FAA

• Kelvin K. Ampofo, AAS-320: Point of contact for CertAlert 24-02; kelvin.k.ampofo@faa.gov; 202-267-6457
• Airport Safety and Operations Division (AAS-300): Issued CertAlert 24-02
• Regional Airport Certification and Safety Inspectors: Primary contact for testing arrangements

EASA

• Innovation Department: innovation@easa.europa.eu
• AI Team: Responsible for AI Roadmap 2.0 and ML concept papers

ICAO

• Aerodrome Panel: Develops Annex 14 amendments
• Safety Management Panel: Develops Annex 19 (SMS) guidance

Singapore CAAS

• enquiry@caas.gov.sg: General enquiries
• Aerodrome Standards Division: Responsible for aerodrome advisory circulars

Industry Organizations

• PAVE Coalition: pavecampaign.org
• SAE AVSC: avsc.sae-itc.com
• ACI World: aci.aero
• IATA Ground Operations: iata.org/ground-ops
• ACRP: trb.org/ACRP
• EUROCAE: eurocae.net
• RTCA: rtca.org

---

This document reflects research conducted through March 2026. Regulatory landscapes change rapidly. Verify current status of all guidance documents, rulemaking activities, and standards development timelines before making strategic decisions.