White Paper: Drone-Based Track Detection to Support Mainline ATI 

Executive Brief 

The restoration of the Federal Railroad Administration’s Automated Track Inspections (ATI) waiver framework creates an opportunity to deploy new sensing technologies that improve safety outcomes without displacing established inspection authorities. One of the most promising support technologies is aerial inspection using unmanned aircraft systems. 

This paper presents a regulatory-aware framework for using drone-based gauge and geometry risk screening to supplement (rather than replace) geometry cars, hi-rail inspections, and FRA Part 213 compliance processes.  

The objective is not automation of compliance, but improved allocation of human attention, measurement assets, and maintenance resources to the locations where risk is emerging. 

Regulatory Framing and ATI Alignment 

FRA and ATI Principles 

Under current FRA regulations and restored ATI waivers: 

• Automated and advanced technologies may supplement, but not replace, required inspections 

• ATI systems are intended to identify conditions humans may miss and improve safety outcomes 

• Final compliance determinations remain grounded in accepted inspection systems, including: 

• Geometry cars and TGMS platforms 

• Hi-rail and walking visual inspections 

• FRA Part 213 geometry limits 

Any drone-based gauge or geometry capability must therefore be positioned as an upstream screening and decision-support layer, not as a regulatory measurement authority. 

Positioning of Drone-Based Gauge Measurement 

Within the ATI ecosystem, drone-based gauge measurement is best characterized as: 

• Capable of detecting relative change and anomaly patterns 

• Useful for early identification of emerging risk 

• Effective at prioritizing where certified measurement assets should be deployed 

• Additional data reference supporting ATI techniques for waiver submissions 

It is explicitly not positioned as: 

• A direct replacement for TGMS/geometry cars or manual inspections 

• A direct source of FRA-reportable defects 

Role of Drone-Based Gauge Screening in the Inspection Stack 

What the Technology Reliably Provides 

Using single-camera RGB imagery and trained AI models, aerial systems can reliably detect: 

• Rail head divergence and convergence trends 

• Abnormal perspective widening indicative of gauge degradation 

• Tie plate, skew and cutting patterns 

• Fastener migration or loss 

• Ballast shoulder loss and void development near the gauge face 

These detections are probabilistic and comparative in nature, producing risk likelihood outputs rather than absolute measurements. 

Acceptable Outputs Under ATI 

An example of an ATI-aligned output would be: 

“Segment MP 123.4–123.9 exhibits a statistically significant widening trend relative to baseline imagery over the prior 90 days. TGMS or hi-rail validation recommended.” 

This type of output enhances, rather than competes with, certified inspection systems. 

Near-Term Operational Concept: Pre-Inspection Risk Screening 

Aerial Screening Layer 

In the near term, fixed-wing or quadcopter drones operate as a pre-inspection screening layer: 

• Corridor or area flights conducted at approximately 30–60 feet AGL 

• Monocular AI models evaluate imagery for gauge- and geometry-related anomalies 

• Flagged locations are tagged with milepost ranges, confidence scores, and trend context 

The output is a heat map of relative geometry risk, not a determination of compliance. 

Optimization of Existing Assets 

Drone-derived risk screening directly improves the efficiency of established programs: 

• Geometry cars are routed or scheduled to validate the highest-risk segments 

• Hi-rail inspections focus walking effort where anomalies are indicated 

• Low-risk mileage receives reduced attention without reducing compliance 

The result is improved utilization of high-cost measurement assets and skilled labor. 

Sample Future State: Integrated Aerial and Ground Inspection 

Fixed-Wing Aircraft as a Network Alerting Layer 

In a mature future state, fixed-wing unmanned aircraft operate as a persistent, corridor-scale alerting layer across large portions of the network. 

Key characteristics of this layer include: 

• Long-range coverage optimized for efficiency rather than resolution 

• Periodic re-flight to establish baseline and trend comparisons 

• Automated flagging of segments exhibiting deviation from historical norms 

These aircraft do not guide maintenance directly. Instead, they generate prioritized alerts that feed downstream inspection planning. 

Guiding Traditional Measurement Technologies 

Fixed-wing alerts inform decisions such as: 

• Where to deploy geometry cars during upcoming inspection windows 

• Which territories require increased hi-rail presence given current conditions 

• Where ad hoc validation should occur ahead of scheduled runs 

In this model, traditional measurement technologies remain the authority, but are guided by broader, data-driven awareness. 

Empowering the Hi-Rail Inspector with Quadcopter Support 

At the local level, the hi-rail track inspector becomes a more capable, data-enabled operator. 

Using a small, semi-automated, lightweight quadcopter carried on the hi-rail vehicle, the inspector can: 

• Launch rapid, localized aerial inspections at flagged locations 

• Obtain overhead and oblique views of gauge faces, fasteners, and ties 

• Document conditions that are difficult to assess from ground level 

• Compare current conditions against recent aerial baselines 

• Inspect areas of the track without fouling or needing track time 

The quadcopter does not replace walking inspections or manual gauge measurements. It enhances visual context, reduces time spent repositioning, allows pre-inspection of a local area ahead of available track time, and improves documentation quality. 

Ad Hoc and Situational Use Cases 

This local aerial capability is particularly valuable for: 

• Investigating alerts generated by fixed-wing screening 

• Assessing conditions in constrained or unsafe environments 

• Supporting post-weather or post-incident inspections 

• Improving situational awareness before committing crews or equipment 

Governance, Safety, and Regulatory Considerations 

The proposed future state maintains conservative governance principles: 

• Human inspectors retain full authority over safety determinations 

• Manual measurements remain the basis for compliance 

• Drone data is treated as advisory and documented as such 

• Transparency and data retention support FRA evaluation 

By preserving these boundaries, the system aligns with the intent of ATI rather than challenging it. 

Strategic Value to the Railroad 

From an enterprise perspective, the combined aerial and ground-enabled model delivers: 

• Earlier detection of emerging gauge and geometry issues 

• Higher return on investment from geometry cars and TGMS assets 

• Reduced unplanned track time and emergency interventions 

• Improved labor efficiency without workforce displacement 

• Stronger, data-backed safety cases for future ATI evolution 

Conclusion 

Drone-based gauge and geometry screening is not a replacement technology; it is an enabling one. When positioned correctly within the ATI framework, it expands visibility, sharpens prioritization, and strengthens the effectiveness of existing inspection authorities. 

A future state that combines fixed-wing network alerting with quadcopter supported inspections represents a logical evolution of today’s programs. It preserves regulatory integrity while materially improving the railroad’s ability to identify, understand, and manage track risk before failures occur.