TractEasy (TLD + EasyMile) Autonomous Baggage Tractor: Production Deployment Report

Last updated: 2026-03-22

1. Corporate Structure and Background

Joint Venture Formation

TractEasy was formally launched on 27 June 2024 as a joint venture between:

Alvest Group -- parent company providing financial backing and industrial base
- TLD -- global leader in ground support equipment (GSE), 9 factories worldwide, 1,800+ employees. Manufactures the vehicle platform.
- Smart Airport Systems (SAS) -- Alvest sister company providing aviation expertise and airport distribution channels
EasyMile -- established driverless technology provider headquartered in Toulouse, France. Supplies the autonomous driving software stack.

CEO: Richard Reno (former CEO of TLD Americas)

In December 2024, TractEasy launched TractEasy GmbH in Berlin, Germany -- its largest non-French office, reflecting Germany's status as the company's biggest market due to its industrial base and innovation ecosystem.

Product Line

Product	Description	Status
EZTow	Autonomous tow tractor for baggage/cargo towing	Production, most-deployed autonomous tow tractor globally
EZDolly	Autonomous cargo dolly for ULD/pallet transport	Serial production started 2025; first 2 units delivered May 2025
EZFleet	Fleet management and supervision software	Production, deployed with every EasyMile-powered vehicle since 2017

Scale

20+ EZTow units deployed globally across airports and industrial sites
Deployments on 3 continents (Europe, Asia, North America) since 2018
400+ deployments of EasyMile's autonomous driving stack overall, with 600,000+ miles driven

2. Every Airport Deployment

2.1 Toulouse-Blagnac Airport, France

Field	Detail
Operator	Alyzia Group (ground handler)
Airport	Toulouse-Blagnac International Airport (TLS)
Start Date	November 2022 (with onboard supervisor)
Level 4 Date	7 November 2023 (fully driverless, no human on board)
Vehicle	EZTow autonomous tow tractor
Fleet Size	1 unit (pilot)
Route	Aircraft landing positions to baggage hall
Initial Route Length	800 m
Extended Route Length	2,000 m (2 km) -- extended upon L4 transition
Route Elements	Intersections, roundabouts, turning circles, narrow indoor luggage gallery
Weather Tested	Rain, fog, snow
Accidents	Zero

Operational Details:

The EZTow tows luggage from aircraft landing positions to the baggage sorting/claim area.
Initially operated with an onboard safety attendant (Level 3).
Progressed to Level 4 on 7 November 2023: the onboard attendant was removed entirely.
The former onboard attendant now monitors missions remotely from a tablet.
Other ramp workers can use a rear panel on the vehicle to dispatch it to any point on the programmed line.
The extended route (800 m to 2 km) added indoor challenges: narrow luggage gallery trajectories and increased interaction with other traffic.

Partners: Alvest Group, TLD, Smart Airport Systems (vehicle); EasyMile (driverless technology); Alyzia (ground handling operations).

Strategic Purpose: Demonstrate autonomous vehicles can optimize luggage/freight logistics while addressing labor shortages. Part of Alyzia's goal to serve more flights and optimize baggage handling.

2.2 Narita International Airport, Japan

Field	Detail
Operator	Japan Airlines (JAL)
Airport	Narita International Airport (NRT), Tokyo
Initial Pilot	October 31, 2019 -- March 31, 2020
Official Operations Start	March 2021 (JAL announced official introduction)
Level 4 Operations	December 17, 2025
Vehicle	EZTow (TractEasy) -- NOTE: JAL also uses ROBO-HI RoboCar Tractor 25T at Haneda
Current Fleet	2 units at Narita (as of Dec 2025)
Planned Fleet	6 units at Narita by ~April 2026
Route	Terminal 2 Main Building to Satellite Terminal baggage sorting area
Infrastructure Changes	None required

Timeline:

September 2019 -- JAL and Narita International Airport Corporation (NAA) announce pilot program
October 2019 - March 2020 -- Initial pilot: transporting air cargo and check-in baggage within restricted area
March 2021 -- JAL officially introduces autonomous towing tractors, becoming first Japanese carrier to do so
December 15, 2025 -- JAL begins transporting checked baggage between Terminal 2 main building and satellite
December 17, 2025 -- Level 4 autonomous operations formally commence after 6-year partnership with TractEasy

Regulatory Approval: Japan's civil aviation authority (JCAB) authorized Level 4 operations, confirming "technical maturity, safety performance, and operational robustness."

JAL's Broader Autonomous Fleet Vision:

6 units at Narita by April 2026
3 units at Haneda by summer 2026 (using ROBO-HI RoboCar Tractor 25T, not TractEasy)
~50 units across the fleet within 5 years
Expansion to 2-3 additional airports after that
JAL President Mitsuko Tottori publicly committed to this roadmap

Note on Haneda: At Haneda, JAL uses a different vehicle (ROBO-HI RoboCar Tractor 25T, max 30 tons towing, 245 km range). ANA separately operates 3 Toyota Industries 3ATE25 tractors at Haneda. Both commenced December 15, 2025 operations. One JAL operator manages up to 10 vehicles via the ROBO-HI OS platform.

2.3 Changi Airport, Singapore

Field	Detail
Operators	Changi Airport Group (CAG) + SATS (ground handler)
Airport	Singapore Changi Airport (SIN)
Proof of Technology	October 2020
Live Trial Start	August 2021 (Terminal 3 environment)
Full Deployment	January 20, 2026 (live driverless operations)
Current Fleet	2 autonomous tractors (live operations as of Jan 2026)
Next Phase	6 additional tractors later in 2026
Long-term Target	24 vehicles by 2027
Routes	T1-T4 baggage handling areas (current); T2 baggage handling to aircraft stands (upcoming)
Regulatory Body	Civil Aviation Authority of Singapore (CAAS)

Timeline:

October 2020 -- Proof of technology trial begins
August 2021 -- Live operational trial in Terminal 3
~2022-2025 -- Months/years of preparation, testing, and validation; accumulated ~20,000 km, more than 5,000 test trips
January 20, 2026 -- Official launch of fully autonomous driverless operations, officiated by Senior Minister of State for Transport Ms. Sun Xueling

Technical Details at Changi:

Max speed: 15 km/h
Tows up to 4 ULDs (unit loading devices) per trip
10+ sensors and cameras per vehicle
Lidar, HD cameras, GPS, 4G, WiFi connectivity
Vehicle location tracked to within 1 centimeter accuracy
Operates in all conditions: day, night, rain

Supervision:

Remote control center monitors operations
Operators available for immediate intervention if required
Clear autonomous vehicle zone markings painted on airside surfaces
Labels attached to all autonomous vehicles for identification

Funding: CAAS co-funded the project.

Workforce Impact:

Frees airside workers from driving tasks
Workers refocused on "last mile operations, which are more difficult to automate"
SATS' "Hub Handler of the Future" programme integrates automation as core focus
Union collaboration on job redesign programmes

Important Note on Multiple Suppliers: Changi Airport works with both TractEasy/EasyMile and UISEE (Chinese autonomous driving company). UISEE claims the January 2026 fleet as its deployment, having also completed 20,000+ km of accident-free operation and 5,000+ trial runs. UISEE holds ISO 21434, ISO 27001, and Singapore TR68 certifications. The exact vendor split between routes/terminals is not fully clarified in public sources, but both suppliers are confirmed to be operating at Changi.

2.4 Greenville-Spartanburg International Airport (GSP), USA

Field	Detail
Operator	Piedmont Airlines (wholly-owned subsidiary of American Airlines) + GSP Airport District
Airport	Greenville-Spartanburg International Airport (GSP), South Carolina
Start Date	September 4, 2024 (demonstration)
Vehicle	EZTow autonomous tow tractor
Towing Capacity	14 tons
Power	Electric battery
Route Length	~1 mile (1.6 km)

Operations:

Cargo operations: Operates on a 1-mile route in mixed traffic, transporting cargo to a staging area during loading/unloading of cargo aircraft
Baggage handling: Operates on passenger terminal ramp for Piedmont Airlines -- picks up baggage from arriving planes at gates and autonomously transports to baggage claim carousel
Operates in mixed traffic with manned vehicles

Regulatory Context: Notable because FAA CertAlert 24-02 (February 2024) stated autonomous vehicles are "not authorized" at Part 139 airports for airside operations. GSP's deployment operates under specific conditions (see Regulatory section below).

Key Quotes:

GSP CEO: "We are always looking for cutting-edge technology that makes the airport experience more safe, efficient, sustainable and affordable."
Piedmont Director: partnership would "improve our safe, reliable, and efficient ground operations."

2.5 Al Maktoum International Airport (DWC), Dubai, UAE

Field	Detail
Operator	dnata (major global aviation services provider)
Airport	Dubai World Central -- Al Maktoum International Airport (DWC)
Announcement	September 12, 2024
Operational Deployment	July 15, 2025
Vehicle	EZTow autonomous tow tractor
Fleet Size	6 units
Investment	AED 6 million (~US$1.6 million)
Current Autonomy	Level 3 (minimal human oversight)
Planned Upgrade	Level 4 by early 2026
Speed	Up to 15 km/h
Capacity	Up to 4 ULDs per vehicle
Routes	Pre-defined pathways between terminal and aircraft

Regulatory Process:

Over one year of collaborative work between dnata, TractEasy, Dubai Airports, and UAE GCAA
GCAA granted the country's first regulatory approval to trial autonomous baggage handling vehicles
Framework "balances innovation with operational safety"
Framework developed jointly by GCAA, Dubai Airports, and dnata
Described as "a new regulatory framework for autonomous vehicle operations in airside environments"

Strategic Context:

DWC is planned to expand to 260 million annual passengers -- this deployment serves as a testbed
Reassignment of human operators to higher-value tasks
Faster aircraft turnarounds
Reduced human error and safety risks

Key Quotes:

Jaffar Dawood, dnata SVP UAE Airport Operations: "This deployment brings the technology into regular, day-to-day operations"
Richard Reno, TractEasy CEO: "Autonomous GSE adoption is taking off"

2.6 Oslo Gardermoen Airport (OSL), Norway

Field	Detail
Project	AWARD H2020 (EU-funded research project)
Airport	Avinor Oslo Gardermoen Airport
Vehicle	EZTow
Focus	Winter weather testing at Level 4 autonomy
Status	Research/testing (not commercial production deployment)

Test Details:

Operated at Level 4 (fully independent, no human on board) in snow conditions
Tested sensor resilience under Norwegian winter weather
Key challenges identified:
- Snow creating uniform landscapes that reduce contrast and object detection
- Snowflakes distorting perceived location, distance, and angle readings
- Multi-sensor consensus disruption when snow affects different sensor types differently
- Risk of false detections or missed detections
Winter tire grip evaluation for stability/maneuverability on icy surfaces
Taxi Lane Departure (TLD) tests for navigation precision

Project Coordinator Quote (Julien Collier): "These tests are instrumental for the EZTow's airport operations, ensuring seamless integration."

2.7 Amsterdam Schiphol Airport, Netherlands

Field	Detail
Operator	KLM Ground Services
Airport	Amsterdam Schiphol Airport (AMS)
Trial Period	February 2021 (~1 month)
Partners	KLM Ground Services, Smart Airport Systems
Status	Trial/pilot

Operations:

Phase 1: Defined area simulating baggage process
Phase 2: Operational environment, bringing baggage to aircraft
Vehicle loaded in baggage area, then navigated to aircraft stand via fixed route
Vehicle manufactured by TLD, autonomous software by EasyMile

2.8 Munich Airport -- Status Clarification

Munich features in TractEasy's history primarily as the location of Inter Airport 2017, where TLD and EasyMile first publicly announced the TractEasy partnership. The Airside International article from December 2025 references Munich as a planned deployment destination alongside DWC, but no detailed public information exists on an active Munich airport operational deployment as of March 2026. Munich should be considered a planned/upcoming site rather than a confirmed production deployment.

3. Technical Architecture in Production

3.1 Vehicle Platform -- EZTow

Specification	Value
Dimensions (L x W x H)	3,200 mm x 1,940 mm x 2,050 mm
Turning Radius	4,750 mm (4.25 m cited in some sources)
Max Speed (autonomous)	15 km/h (9.3 mph); some sources cite up to 25 km/h unladen
Operational Speed	10 km/h typical at industrial sites
Towing Capacity	14 tons (28,000 lbs) standard; up to 25 tons cited in marketing
Power	Fully electric
Battery Options	Lead-acid or Li-ion
Autonomy Level	SAE Level 4 (High Automation per SAE J3016)

3.2 Sensor Suite

The EZTow carries a comprehensive multi-sensor array:

Sensor Type	Purpose
LiDAR (multiple units)	3D environmental mapping, localization, obstacle detection
Stereo cameras	Visual perception, object classification
Radar	Obstacle detection (robust in adverse weather)
IMU (Inertial Measurement Unit)	Motion tracking, dead reckoning
GPS/GNSS	Positioning (multi-GNSS: GPS + GLONASS)
Wheel encoders	Odometry, distance measurement
V2X on-board units	Vehicle-to-infrastructure (V2I) communication
3G/4G modem	Cellular connectivity for remote monitoring
WiFi	Local connectivity

At Changi Airport, each vehicle carries 10+ sensors and cameras.

LiDAR Supplier: TLD signed a 3-year agreement with Velodyne (announced April 2020) to use Velodyne LiDAR sensors in production. Following the Velodyne-Ouster merger (completed February 2023), the sensors are now supplied under the Ouster brand. Velodyne/Ouster sensors provide:

Surround-view LiDAR with data-rich 3D point clouds
Detection of low-reflectivity objects
Optimized indoor/outdoor performance across varying light conditions
High-resolution 3D perception with broad vertical field of view

3.3 Compute Architecture

EasyMile's autonomous driving stack runs on a dual-computer architecture:

Decision-making computer -- runs high-level path planning and decision-making software
Dedicated safety computer -- runs separate safety-critical software with certified safety PLCs

This separation ensures that even if the decision-making software encounters issues, the safety system operates independently.

Localization:

Centimeter-level positioning accuracy (within 2-3 cm typically; 5 cm minimum guaranteed)
At Changi, accuracy to within 1 centimeter
Uses sensor fusion algorithm combining all sensor data
Advanced mapping with pre-built 3D maps of the environment
Multi-GNSS processing (GPS + GLONASS)
Odometry from wheel encoders
IMU for dead reckoning between GNSS fixes

Navigation Method:

Operates on pre-defined routes (not free-roaming)
Routes are programmed/mapped in advance
V2I (vehicle-to-infrastructure) information exchange for route optimization
No airport infrastructure modifications required

Path Planning:

Evaluates situation and determines safest, most efficient trajectory
Decisions governed by certified safety logic
Three response modes: adapt speed/trajectory, controlled braking, or full stop

3.5 Connectivity

4G/3G cellular for remote monitoring and fleet management
WiFi for high-bandwidth local connectivity
V2X for infrastructure communication
Constant communication with EZFleet supervision system

3.6 Fleet Management -- EZFleet

EZFleet is EasyMile's fleet management platform, deployed with every autonomous vehicle since 2017:

Real-time monitoring: Vehicle positions, assigned routes, ETA, destinations, vehicle status
Dynamic dispatch: Assign missions dynamically or switch between fixed-route and on-demand service
Traffic integration: Real-time traffic information for service optimization
Data analytics: Collection, analysis, and reporting
Multi-vehicle management: Supervise entire fleet from smartphone or tablet
API integration: Integration with existing customer systems (confirmed for EZDolly with Airport Operations Management Systems)
VDA 5050 compatible: Through SYNAOS partnership, enabling interoperability with other AGV suppliers

3.7 EZDolly (Next-Generation Product)

Specification	Value
Base Platform	TLD "TF" transporter
Load Capacity	7 metric tonnes
ULD Support	All major upper/lower deck containers and pallets (full 96" width)
Drive	iBS-powered electric driveline
Maneuverability	360-degree
Manual Station	None (fully autonomous only, no driver position)
Operation	24/7
Production Start	2025

4. Operational Procedures

4.1 Pre-Deployment: Getting Airport Authority Approval

There is no dedicated international legal framework for autonomous vehicles at airports. Each deployment requires bespoke approval through a multi-stakeholder process:

Typical Approval Process:

Stakeholder alignment -- Airport authority, ground handler, AV provider, and national civil aviation authority collaborate (typically 1+ year)
Risk assessment -- Site-specific safety assessment covering the Operational Design Domain (ODD)
Controlled testing -- Initial testing in closed/controlled areas
Progressive autonomy -- Graduated approach:
- Phase 1: Manual operation with safety driver
- Phase 2: Autonomous with onboard safety attendant (Level 3)
- Phase 3: Fully autonomous, no human on board (Level 4) with remote supervision
Operational approval -- Civil aviation authority grants permission based on demonstrated safety

Country-Specific Experiences:

Country	Authority	Approach
Japan	JCAB	Authorized L4 after structured testing/validation over 6 years
Singapore	CAAS	Co-funded project; approved after ~1 year of trials and 5,000+ test trips
UAE	GCAA	Created country's first regulatory framework for autonomous airside vehicles; 1+ year development
USA	FAA	CertAlert 24-02 (Feb 2024) states AGVS "not authorized" at Part 139 airports; supports controlled testing only
France	DGAC	Permitted graduated deployment at Toulouse
EU	EASA	Advocates ICAO-level standardization; March 2025 regulations create foundation for AGHE integration

4.2 Safety Case and Risk Assessment

Standards Compliance:

ISO 13849-1 -- Safety-related parts of control systems: TractEasy achieves Performance Level PL d (second-highest level)
CE marking -- Products meet applicable European safety directives
ISO/EN safety standards -- Certified compliance
Rigorous risk assessments with comprehensive technical documentation

Safety Analysis Tools:

EasyMile uses Ansys medini analyze for functional safety analysis
Implements: HAZOP, HARA, FHA, FTA, FME(C)A, FMEDA methods
Safety analysis customized into a single model covering both passenger transport and material handling platforms
SOTIF (Safety of the Intended Functionality) analysis for sensor/algorithm safety in absence of system failure

Safety Architecture:

Redundant braking paths (fail-safe braking system)
Certified safety PLCs
Continuous self-diagnostics
Dual-computer separation (decision vs. safety)
Multiple layers of redundant systems
360-degree sensor coverage

4.3 Driving Modes

The EZTow supports four modes:

Automated mode -- Full autonomous operation
Manual approach mode -- Transitional mode for docking/positioning
Manual mode -- Human driver control
Inching mode -- Very slow speed for fine positioning

4.4 Operational Workflow (Typical Baggage Towing)

Baggage loaded into ULDs at baggage handling area or from arriving aircraft
ULDs attached to EZTow (up to 4 ULDs)
Mission dispatched via EZFleet (tablet, rear panel, or automated dispatch)
EZTow navigates pre-defined route autonomously
Remote supervisor monitors via tablet/control center
EZTow arrives at destination (baggage hall, aircraft stand, or carousel)
ULDs detached and processed

4.5 Supervision Model

Level 3 Operations (with onboard attendant):

Safety attendant rides on vehicle
Can take manual control at any time
Monitors vehicle behavior and surroundings

Level 4 Operations (no human on board):

Former onboard attendant monitors from a tablet (remote)
Site Control Center provides remote monitoring and supervision
Real-time communication with vehicle fleet
Vehicle Operators and Field Operators track/control fleet from smartphone or tablet
Remote control capabilities: rearming, starting/resuming missions, recording logs
Other workers dispatch vehicle via rear panel on the vehicle itself

Teleoperation Technology:

EasyMile uses DriveU.auto for remote supervision/piloting
Patented cellular bonding technology -- fuses multiple cellular networks for reliability
Dynamic video encoding for optimized bandwidth
Industry's lowest latency for video supervision
Multiple cellular modems per vehicle (single network insufficient for safety)

5. Edge Cases and Failure Handling

5.1 How the System Handles Edge Cases

Obstacle Detection Response (3 tiers):

Adapt -- Adjust speed and trajectory to avoid obstacle
Brake -- Controlled braking if path cannot be adjusted
Stop -- Full emergency stop if situation requires

Known Edge Cases in Airport Environment:

Intersections with crossing traffic
Roundabouts (tested at Toulouse)
Turning circles with tight clearances
Narrow indoor areas (luggage gallery at Toulouse)
Mixed traffic with manned vehicles, pedestrians, aircraft
Varying light conditions (day, night, hangar interiors)

5.2 Winter/Weather Edge Cases

Identified Challenges from Oslo Testing:

Snow creating uniform landscapes that reduce contrast for computer vision
Snowflakes distorting LiDAR readings (location, distance, angle)
Multi-sensor consensus failure when snow affects sensors differently
Risk of both false positive (phantom obstacle) and false negative (missed real obstacle) detections
Reduced tire grip on icy surfaces

Mitigation Approaches:

Sensor fusion algorithm cross-validates readings from multiple sensor types
Winter tire grip evaluation
Continuous self-diagnostics detect when sensor confidence degrades
Vehicle defaults to safe stop when confidence is insufficient

5.3 Human Takeover Procedures

Remote operator in control center can intervene immediately (confirmed at Changi)
Workers on the ground can use rear panel to stop/redirect vehicle
Manual mode available for immediate human control
Vehicle enters safe stop state when any anomaly detected

5.4 Safety Record

Zero accidents across all deployments (stated by TractEasy/EasyMile as of latest reports)
Zero collisions in EasyMile's track record (company-wide claim)
>95% autonomous mission success rate
Changi: 20,000+ km of accident-free operation
No public incident reports found in FAA (NTSB), EASA, or other aviation safety databases

6. Performance Metrics

6.1 Published Metrics

Metric	Value	Source/Context
Autonomous mission success	>95%	TractEasy global
Accident rate	0 (zero accidents)	All deployments
Localization accuracy	1-5 cm	Varies by site; 1 cm at Changi
Operating hours	24/7 capability	Confirmed at industrial sites
Test trips at Changi	5,000+	Pre-deployment validation
Distance at Changi	20,000+ km	Accident-free operation
Max ULDs per trip	4	Standard at most airports
Max tow train length	16 m (3 trailers)	BMW Dingolfing

6.2 Metrics NOT Publicly Available

The following metrics have not been disclosed in public sources:

Bags per hour throughput
Specific uptime percentage (only "24/7 capability" stated)
Direct comparison with manual operations (quantitative)
Mean time between failures (MTBF)
Cost per bag moved vs. manual
Turnaround time improvement data
Battery range/endurance per charge cycle

6.3 Qualitative Performance Indicators

At BMW Dingolfing, the plant progressively increased EZTow from 1 shift to 3 shifts over the course of a year, indicating the system proved reliability before scaling
Toulouse extended route from 800 m to 2 km after demonstrating capability
JAL's 6-year progression from pilot to L4 production reflects conservative, validated scaling
dnata describes the technology as ready for "regular, day-to-day operations"

7. Scaling Journey

7.1 Single Vehicle to Fleet

The typical scaling pattern observed across deployments:

Phase 1: Proof of Concept (3-12 months)

Single vehicle in controlled area
Simulated operations
Example: Schiphol (Feb 2021, 1-month trial)

Phase 2: Live Operations with Supervision (6-24 months)

Single vehicle on live airside with onboard attendant
Serving real flights
Example: Toulouse Nov 2022 (L3, onboard supervisor)

Phase 3: Full Autonomy, Single Vehicle (6-12 months)

Remove onboard human, remote supervision only
Extended routes
Example: Toulouse Nov 2023 (L4, route extended 800 m to 2 km)

Phase 4: Fleet Deployment (12+ months)

Multiple vehicles on multiple routes
Fleet management coordination
Example: Changi (2 vehicles on T1-T4 route, expanding to 8, then 24 by 2027)

7.2 What Changed During Scaling

Technical Evolution:

Sensor fusion algorithms refined based on real-world data
Route complexity increased (simple point-to-point to routes with intersections, roundabouts, indoor/outdoor transitions)
Weather resilience improved through AWARD H2020 testing

Operational Evolution:

Onboard attendant role transformed to remote supervisor role
Fleet management software matured (EZFleet deployed since 2017)
Integration with SYNAOS for multi-vendor AGV orchestration (announced 2025)

Known Scaling Challenges:

Integrating multiple AGV suppliers into a single fleet management system (addressed by SYNAOS partnership and VDA 5050 standard)
Each airport requires bespoke regulatory approval -- no global standard
High capital expenditure vs. traditional manned vehicles
Richard Reno: "Scaling up the deployment is the next challenge in aviation"
Only providers with "highest level of safety will progress from proof-of-concept to widescale adoption"

7.3 Industrial Scaling Reference: BMW Dingolfing

The BMW Group Plant Dingolfing deployment provides the clearest scaling trajectory:

Use case: Towing PHS-hardened sheet metal parts in outdoor manufacturing areas
Traction: 14 tons, three trailers, 16+ m combined length, 10 km/h
Scaling: 1 shift to 3 shifts over 1 year
Key insight: Plant increased tractor capacity only after it "proved its reliability and performance with smooth operation and the ability to handle harsh weather"
Lesson: Scaling principle is "the greater the number of vehicles at a single location, the higher the potential for synergies, better support structures, maintenance capacity, and return on investment"

8. Regulatory Approvals

8.1 Standards and Certifications

Standard	Status
ISO 13849-1 (Safety of machinery)	Compliant, Performance Level PL d
CE Marking	Products CE-marked per EU safety directives
SAE J3016	Level 4 High Automation classification
ISO/EN safety standards	Certified compliance
Functional Safety Analysis	HAZOP, HARA, FHA, FTA, FMEA via Ansys medini analyze

8.2 Regulatory Landscape

No unified international framework exists. The regulatory environment is fragmented:

EASA (Europe):

Has received industry requests for autonomous GSE framework
Advocates for ICAO-level standardization rather than regional rules
Julia Egerer (EASA head of aerodrome safety): "The best way forward is to discuss this technology at an ICAO level"
March 2025 EASA regulations create regulatory foundation for scalable Automated Ground Handling Equipment (AGHE) integration across European operations

FAA (USA):

CertAlert 24-02 (February 15, 2024): AGVS testing/deployment/operation "not authorized" at Part 139 airports
Supports testing in controlled environments: remote areas, landside, or movement areas closed to aircraft operations
Acknowledges existing safety requirements "were not originally developed with autonomous vehicles in mind"
Working on developing standards and guidance
Emerging Entrants Bulletin 25-02 (2025) provides updated testing guidance

JCAB (Japan):

Authorized Level 4 operations at Narita (December 2025)
JCAB Director-General Koichi Miyazawa emphasized autonomous airport technology significance

CAAS (Singapore):

Co-funded autonomous tractor project
Approved deployment after rigorous trials
Collaborative approach with unions for workforce transition

GCAA (UAE):

Created country's first regulatory framework for autonomous airside vehicles
Developed jointly with Dubai Airports and dnata over 1+ year

IATA:

Developed recommended practices for testing and implementing autonomous GSE
Working on use cases and end-to-end automation processes
Enhanced GSE Recognition Program: 98 fleets registered as of May 2025; mandatory declarations at ISAGO-accredited locations

8.3 Certification Timeline Observations

Narita: 6 years from first pilot (2019) to Level 4 authorization (2025)
Changi: ~5 years from proof of technology (2020) to full deployment (2026)
Toulouse: ~1 year from supervised operations (Nov 2022) to Level 4 (Nov 2023)
DWC: 1+ year regulatory framework development before deployment
Common theme: Years of incremental validation, not rapid deployment

9. Lessons Learned and Public Statements

9.1 From TractEasy/EasyMile

Richard Reno, CEO TractEasy:

"Autonomous GSE adoption is taking off"
"With technology evolution, building pools of autonomous vehicle expertise and growing understanding of and capability to address operation environment barriers in the airport market, autonomous tow tractor deployments are set to take off"
"For pioneering airports and airlines implementing the technology, benefits are being achieved or exceeded, and scaling plans have graduated from the white board to the board room for approval"
"Only solutions and providers with the highest level of safety will progress from proof-of-concept to widescale adoption"

Marion Ferre, EZTow Deployment Project Manager (EasyMile):

Detailed that L4 achievement means the vehicle is "truly capable of maneuvering and navigating complex scenarios on their own" while being "commercially viable"
Highlighted that removing the onboard human unlocks "cost and time efficiency, scalability and flexibility"

9.2 From Airport Operators and Airlines

JAL President Mitsuko Tottori:

Committed to ~50 autonomous units within 5 years, expansion to 2-3 additional airports

SATS (Changi) -- Kuah Boon Kiam, SVP Apron Services:

"This initiative supports SATS' Hub Handler of the Future programme, where the integration of automation into our airside operations is a core focus to enhance safety, boost turnaround efficiency, and uplift service quality"

dnata -- Jaffar Dawood, SVP UAE Airport Operations:

"This deployment brings the technology into regular, day-to-day operations"
Automation could be vital for "smarter, safer and more resilient infrastructure"

9.3 From Regulators and Industry Bodies

EASA (Julia Egerer):

"Industry players want standardisation because if there is a common framework, they will know what to expect"
Warned that regional variations would create "chaos"

FAA:

Existing safety requirements "were not originally developed with autonomous vehicles in mind"
Airport operating areas present "significantly different hazards and complexities due to higher speed aircraft operations and congestion"

JCAB (Koichi Miyazawa, Director-General):

Emphasized that expansion requires "cross-company and public-private collaboration"

ASA World Director Fabio Gamba:

Acknowledged cost barriers: autonomous equipment "significantly exceeds traditional vehicle prices"
"We have always been a labour-intensive industry and I don't think that is going to change in the coming years"

9.4 Key Lessons Synthesized

Incremental trust-building is essential: Every deployment follows a years-long graduated pathway from controlled testing to supervised operations to full autonomy
Regulatory fragmentation is the biggest barrier: No global standard exists; each country/airport requires bespoke approval
The technology works, but scaling is the challenge: Zero accidents, >95% mission success -- the bottleneck is not technology maturity but regulatory, operational, and commercial readiness
Labor shortages drive adoption: Nearly every deployment cites workforce challenges as a primary motivator
No infrastructure changes needed: The EZTow operates on existing airport infrastructure, which is a critical selling point
Fleet economics improve with scale: More vehicles per site = better ROI, support structures, and maintenance capacity
Weather remains an active research area: Snow, fog, and rain affect sensor performance; the AWARD H2020 project specifically targets this gap
Multi-vendor orchestration is emerging: The SYNAOS partnership signals the industry is moving toward fleet management platforms that coordinate vehicles from multiple suppliers

10. Integration with Airport Systems

10.1 Current Integration Level

Direct Integrations Confirmed:

EZFleet fleet management supervises all autonomous vehicles on site
V2X/V2I communication enables information exchange with airport infrastructure
API integration to customer Airport Operations Management Systems (confirmed for EZDolly)
SYNAOS partnership enables integration with VDA 5050-compliant systems, allowing EZTow to operate alongside other AGVs from different manufacturers

Integration Approach:

EZFleet serves as the middleware between autonomous vehicles and airport operations
Capable of integrating real-time traffic information
Remote monitoring from Site Control Center with fleet-wide visibility

10.2 What is NOT Publicly Confirmed

There is no public evidence of deep integration with:

Baggage Handling Systems (BHS) -- automated handoff between BHS conveyors and EZTow
Flight Information Display Systems (FIDS) -- automatic mission generation based on flight schedules
Airport Operational Database (AODB) -- real-time flight status triggering autonomous vehicle dispatch
A-CDM (Airport Collaborative Decision Making) systems

The current operational model appears to rely on human-dispatched or pre-scheduled missions rather than fully automated system-to-system integration with core airport IT platforms. The EZDolly's announced API integration with Airport Operations Management Systems suggests this capability is under development.

10.3 Future Integration Direction

The SYNAOS partnership (announced March 2025) is significant because SYNAOS provides:

Intralogistics platform built on cloud technology
VDA 5050 standard compliance
Orchestration of mobile robots, IoT devices, manual vehicles, and human workers
Integration capability that could bridge to airport-specific systems

11. Night Operations, Weather Handling, and Adverse Conditions

11.1 Night Operations

Confirmed capability: Changi Airport explicitly states vehicles operate in "all conditions -- day, night and rain"
The EZTow's sensor suite (LiDAR, radar, cameras) is designed for day/night operation
LiDAR and radar are inherently not dependent on ambient light
Cameras supplemented by vehicle-mounted lighting
24/7 operation demonstrated at BMW Dingolfing (3-shift operations)

11.2 Rain Operations

Tested and demonstrated at Toulouse (rain, fog, snow)
Confirmed operational at Changi (tropical climate with heavy rainfall)
Competitor reference airside AV stack developed rain-sensing algorithms effective in rainfall exceeding 50 mm/hour -- TractEasy's specific rainfall threshold not published

11.3 Fog and Low Visibility

Tested at Toulouse in fog conditions
Multi-sensor fusion provides redundancy: when camera/LiDAR visibility degrades, radar maintains detection capability
System designed to degrade gracefully -- reduces speed or stops when confidence drops

11.4 Snow and Ice (Winter Operations)

AWARD H2020 Testing at Oslo Gardermoen:

Achieved Level 4 operation in snow
Critical findings:
- Snow coverage creates uniform terrain that challenges computer vision
- Snowflakes can cause LiDAR to produce distorted readings
- Different sensors degraded in different ways, challenging fusion consensus
- Winter tire grip testing showed the vehicle maintained stability on ice
No detailed temperature range or snow depth thresholds published

11.5 De-icing Periods

No public information exists specifically addressing:

Operations during active de-icing procedures on the apron
Interaction with de-icing fluid on road surfaces
Coordination with de-icing vehicle movements
Impact of de-icing chemicals on sensors

This represents a gap in publicly available operational data.

11.6 Operational Design Domain (ODD) Summary

Based on available information, the EZTow's ODD includes:

Parameter	Specification
Speed	Up to 15 km/h autonomous; 20-25 km/h unladen/manual
Environment	Indoor, outdoor, road surfaces
Weather	Rain, fog, snow (tested); specific limits not published
Time of day	24/7 (day and night)
Traffic	Mixed traffic with manned vehicles and pedestrians
Infrastructure	Intersections, roundabouts, turning circles, narrow corridors
Routes	Pre-defined only (not free-roaming)
Localization	Requires pre-built map; GNSS availability recommended

12. Competitive Landscape

TractEasy is not alone in the autonomous airport towing market. Key competitors identified:

Company	Product	Notable Deployments
UISEE (China)	Autonomous tractor	Changi Airport (Jan 2026), 21 airports globally
reference airside AV stack (UK)	autonomous baggage/cargo tug, autonomous baggage dolly, autonomous cargo vehicle	Stuttgart, Inverness, Cincinnati/NKY, 4-unit fleet (May 2024)
Charlatte	Autonomous baggage tractor	Demonstrated at GSE Expo Europe 2024
Toyota Industries	3ATE25 EV tractor	ANA at Haneda (3 units, Dec 2025)
ROBO-HI	RoboCar Tractor 25T	JAL at Haneda (1 unit, Dec 2025)
Fraport/various	Testing program	Frankfurt Airport (first test run March 2024)

TractEasy claims to be the "most-deployed autonomous tow tractor globally" and appears to maintain that position in the airport segment, though UISEE claims operations at 21 airports worldwide.

13. Summary: Deployment Status Matrix

Airport	Country	Operator	Status	Fleet	Autonomy Level	Year Started
Toulouse-Blagnac	France	Alyzia	Production pilot	1	L4 (Nov 2023)	2022
Narita	Japan	JAL	Production	2 (6 by Apr 2026)	L4 (Dec 2025)	2019
Changi	Singapore	CAG/SATS	Production	2 (24 by 2027)	L4 (Jan 2026)	2020
Greenville-Spartanburg	USA	Piedmont/AA	Demonstration	1	L3/L4	2024
Al Maktoum (DWC)	UAE	dnata	Production	6	L3 (L4 early 2026)	2024/2025
Oslo Gardermoen	Norway	AWARD H2020	Research/testing	1	L4 (testing)	Research
Schiphol	Netherlands	KLM GS	Trial (completed)	1	Supervised	2021
Munich	Germany	TBD	Planned	TBD	TBD	Announced

SLAM Methods

Methods

TractEasy (TLD + EasyMile) Autonomous Baggage Tractor: Production Deployment Report ​

1. Corporate Structure and Background ​

Joint Venture Formation ​

Product Line ​

Scale ​

2. Every Airport Deployment ​

2.1 Toulouse-Blagnac Airport, France ​

2.2 Narita International Airport, Japan ​

2.3 Changi Airport, Singapore ​

2.4 Greenville-Spartanburg International Airport (GSP), USA ​

2.5 Al Maktoum International Airport (DWC), Dubai, UAE ​

2.6 Oslo Gardermoen Airport (OSL), Norway ​

2.7 Amsterdam Schiphol Airport, Netherlands ​

2.8 Munich Airport -- Status Clarification ​

3. Technical Architecture in Production ​

3.1 Vehicle Platform -- EZTow ​

3.2 Sensor Suite ​

3.3 Compute Architecture ​

3.4 Navigation System ​

3.5 Connectivity ​

3.6 Fleet Management -- EZFleet ​

3.7 EZDolly (Next-Generation Product) ​

4. Operational Procedures ​

4.1 Pre-Deployment: Getting Airport Authority Approval ​

4.2 Safety Case and Risk Assessment ​

4.3 Driving Modes ​

4.4 Operational Workflow (Typical Baggage Towing) ​

4.5 Supervision Model ​

5. Edge Cases and Failure Handling ​