Innovation Brief
The Innovation Brief defines the technical foundation of this project. It documents the problem being addressed, the proposed system architecture, prior art research, and the evaluation criteria used to validate the design. This phase forces clarity — what the system is, why it matters, and how it will be measured.
Subway Sentinel
A working brief for a rail-mounted robotic GPR inspection platform. This page captures the project foundation, research basis, innovation claim, MVP scope, evaluation criteria, and engineering tasks guiding the design and validation work.
Technical Field
This project involves the technical fields of industrial six-axis robotics and automation, rail-based mechanical systems, non-destructive inspection technologies, and safety-focused systems engineering, with a focus on robotic ground-penetrating radar (GPR) inspection.
Background Information
Subway Sentinel emerged from a lifelong connection to New York City, its subway system, and the workforce that built and maintains it. Daily reliance on the subway made its importance unmistakable, while recurring delays and service disruptions revealed the challenges created by aging tunnel infrastructure. The idea was further formed by close, long-term exposure to the work of a family member who was a sandhog, giving a firsthand window into the realities of underground tunnel work.
Sandhogs are highly specialized tunnel workers who operate far below the streets of New York City, maintaining the infrastructure that keeps the subway running. Their work demands exceptional technical skill and places them in confined, high-risk environments shaped by unstable ground, water intrusion, and deteriorating structures. Those realities point to the need for modern inspection tools that reduce human exposure while supporting the expertise of the crews who know the system best.
Combined with my professional background in industrial robotics and my interest in applying six-axis robotic systems to meaningful real-world problems, these experiences led to the concept for a rail-mounted robotic inspection platform. Subway Sentinel is designed to support transit inspection teams by reducing human exposure during tunnel inspections while improving the consistency and safety of long-term infrastructure assessment.
Prior Art Research
The prior art review informed the project direction by comparing truck-based systems, hyrail systems, mobile tunnel robots, inspection vehicles, and rail-mounted concepts. Most existing solutions are limited by fixed sensor placement, slow stop-and-scan operation, surface-only inspection, or platforms not designed around an industrial robotic arm. Subway Sentinel is different because it centers on a powered, rail-integrated platform with a full-size industrial robot designed for repeatable, high-control tunnel inspection. GPR is the MVP payload, with a clear path toward modular sensor expansion.
Description: Tunnel inspection system built on an eight-ton truck carrying an industrial robotic manipulator with hydraulic hammers for impact-sound inspection.
Difference: Subway Sentinel uses a rail-mounted six-axis industrial robotic arm and a more rigid rail-aligned base to support more controlled inspection positioning.
Description: Commercial tunnel inspection platform using a GPR sensor mounted on a telescopic arm attached to a road/rail-capable truck.
Difference: Subway Sentinel uses a six-axis robotic arm for greater articulation, repeatability, and future multi-modal inspection capability.
Description: Vehicle-mounted tunnel inspection system combining a mobile lift, robotic arm, computer vision, and ultrasonic sensors for visible lining defects.
Difference: Subway Sentinel replaces the lift-and-crane style approach with a rail platform and industrial arm intended for repeatable corridor inspection.
Description: Truck-mounted tunnel inspection robot using an articulated arm and wall-adhered sensor modules for concrete tunnel lining inspection.
Difference: Subway Sentinel shifts the deployment method from truck-based access to a rail-mounted industrial robotic platform suited to subway operating geometry.
Description: Mobile robotic platform with a KUKA industrial arm for repetitive construction tasks such as drilling and installation.
Difference: Subway Sentinel applies industrial robotic manipulation to rail-mounted inspection and GPR positioning instead of construction automation.
Description: Camera-based subway tunnel inspection system using multiple vision sensors and a neural network approach to detect surface defects.
Difference: Subway Sentinel goes beyond surface-level visual inspection by using GPR to target subsurface defects, voids, and hidden structural issues.
Description: GPR tunnel inspection vehicle that uses rapid scanning and additional sensors to collect data in subway tunnels.
Difference: Subway Sentinel uses a rail-mounted industrial robot arm to move the GPR with tighter control and repeatability instead of relying on fixed sensor geometry.
Description: Mobile tunnel inspection robot using a robotic arm, laser sensors, GPR, and a control algorithm to maintain surface tracking.
Difference: Subway Sentinel emphasizes mechanical stability and rail alignment first, reducing dependence on complex control compensation for the MVP concept.
Description: Autonomous or push-type rail inspection concept using laser scanning and positioning equipment for tunnel health monitoring.
Difference: Subway Sentinel is a powered rail-integrated robotic platform intended to support larger payloads, GPR, and modular inspection passes.
Description: Railcar-mounted robotic system for detecting and tightening loose rail fastener bolts using torque sensing, vision, and remote monitoring.
Difference: Subway Sentinel is sensor-first and focused on tunnel infrastructure inspection, not rail fastener maintenance.
Description: Pickup-truck-mounted six-axis robotic arm used to measure tunnel geometry and deformation during construction.
Difference: Subway Sentinel is designed around a rail-aligned base, robotic GPR positioning, and operational tunnel inspection rather than slow construction-phase surveying.
Description: Mobile tunnel inspection robot using multiple articulated arms to position sensors against tunnel linings for defect detection.
Difference: Subway Sentinel uses one industrial-scale arm on a powered rail platform to carry GPR and modular long-range sensors over larger tunnel sections.
Description: Rail-car mounted system with compact telescoping arms that hold radar sensors near tunnel linings for subsurface inspection.
Difference: Subway Sentinel uses a six-axis industrial arm for more flexible positioning, future tool changes, and broader multi-sensor expansion.
Description: Rail-mounted inspection robot combining camera inspection, positioning, laser SLAM, and a robotic arm for rail vehicle and track-adjacent inspection.
Difference: Subway Sentinel is purpose-built for tunnel infrastructure inspection using GPR mounted to an industrial robotic arm to detect subsurface risks.
Project Description
Subway Sentinel offers a new approach to inspecting aging subway tunnels by using a rail-mounted six-axis industrial robot. This system is developed for transportation infrastructure inspection and non-destructive evaluation, with a strong focus on making underground work safer and more consistent. It can handle tasks such as scanning tunnel walls and crowns and running inspection routines with ground-penetrating radar. Future upgrades may include sensors such as LiDAR and thermal imaging. The system is intended for transit engineers, inspection teams, and maintenance planners who need reliable, repeatable data without placing people in unnecessary tunnel hazards.
Innovation Claim
The core innovation is the inspection platform itself. Subway Sentinel uses a rail-mounted, six-axis industrial robotic arm designed specifically for subway tunnel environments. This platform provides speed, repeatability, and accuracy that mobile or vehicle-based systems can struggle to achieve. The system also supports modular inspection tools, allowing future expansion into LiDAR, thermal imaging, low-light imaging, or other inspection payloads. The result is safer inspections and higher-quality data to support predictive maintenance planning.
Usage Scenario
Scenario: Preventive Monitoring
Subway Sentinel can serve as part of a routine preventive monitoring program on active subway lines. During scheduled maintenance windows, the rail-mounted robot conducts consistent inspections with ground-penetrating radar and other sensors, identifying early signs of water intrusion, voids, displacement, or structural wear. By analyzing inspection data over time, transit agencies can spot potential problems before they cause disruptions, supporting smarter maintenance planning and safer long-term infrastructure decisions.
Evaluation Criteria
The system will be evaluated using Yes/No criteria organized into five sections. These criteria are designed to keep the MVP measurable, realistic, and transparent.
A. Railcar Platform Innovation - Geometry & Interfaces (Yes/No)
- Rail Interface: Is the railcar modeled with steel rail wheels that seat on the rail head?
- Track Gauge: Is the wheel spacing designed to match the documented track gauge assumption?
- Robot Mounting Plate: Is the robot mounting plate modeled with a defined bolt pattern that matches the selected ABB robot base?
- Wheelbase and Mount Height: Are wheelbase length/width and robot mounting height defined for stability evaluation?
- Structural Continuity: Is the frame modeled as a continuous structure connecting the robot mount to the wheel/axle supports?
B. Loads, Strength, Stiffness, and Stability (Yes/No)
- ABB Foundation Loads Identified: Are ABB “Loads on foundation” values captured for Endurance and Max load cases?
- Load Combination Assumptions Documented: Are conservative load assumptions documented?
- Foundation Requirements Considered: Are ABB foundation requirements acknowledged at an MVP level?
- Loads Applied Correctly: Are both forces and moments applied at the mounting plate location?
- Static Strength Check Completed: Has static structural analysis been completed using ABB foundation loads?
- Overbuilt Safety Margin: Does the design meet static Factor of Safety ≥ 3.0?
- Deflection Limit Met: Is predicted displacement at the robot mounting plate ≤ 1.0 mm?
- Anti-Tip Margin Verified: Does the railcar maintain a tip safety margin ≥ 1.5×?
- Low CG / Ballast Strategy Defined: Is a ballast or low-center-of-gravity strategy defined?
C. EOAT + GPR Mounting (Yes/No)
- EOAT Interface Defined: Is a custom EOAT modeled from the ABB wrist flange to the GPR mounting interface?
- EOAT Build Definition: Does the EOAT package include documented dimensions, mounting method, and fastener details?
- Payload Data Set for Simulation: Is the combined EOAT + GPR mass and center of gravity entered as RobotStudio load data?
D. Motion Study in RobotStudio (Yes/No)
- Reach and Configuration: Can the robot reach the defined tunnel inspection region without joint limit violations?
- Singularity Management: Can the robot complete the inspection sweep without singularities, or are mitigations documented?
- Collision Avoidance: Does the motion avoid collisions with the tunnel envelope, railcar, and EOAT/GPR assembly?
- Tool Orientation Control: Is the EOAT/GPR kept at a defined orientation relative to the tunnel surface?
E. MVP Documentation Outputs (Yes/No)
- Core Concept Diagrams Included: Are the system block diagram, railcar concept, and EOAT concept included?
- RobotStudio Proof Video: Is a short exported demo video included showing at least one inspection cycle?
- MVP Limits Declared: Are MVP limitations explicitly listed?
Goals and Tasks Associated with the Project
The objectives below define the measurable deliverables for the MVP, along with the engineering tasks required to complete the work.
Objective 1: Establish Engineering Requirements and Design Assumptions
- Document assumed track gauge and tunnel clearance envelope.
- Define the inspection region and MVP simplifications.
- Set success metrics: FoS ≥ 3.0, mount plate deflection ≤ 1.0 mm, anti-tip margin ≥ 1.5×, and no joint limit violations.
Objective 2: Design a Structurally Credible Railcar Platform
- Model wheelset and rail interface.
- Model railcar frame and robot mounting plate.
- Extract ABB foundation forces and moments from the selected robot manual.
Objective 3: Validate Structural Integrity and Stability
- Run static structural analysis using ABB foundation loads.
- Record stress, displacement, and factor of safety.
- Complete anti-tip stability calculation and iterate the design if needed.
Objective 4: Design EOAT and Define Realistic Robot Load Data
- Model simplified GPR form factor and EOAT adapter.
- Define mounting method, fasteners, mass, and center of gravity.
- Enter EOAT + GPR payload data into RobotStudio.
Objective 5: Demonstrate Motion Feasibility in RobotStudio
- Build RobotStudio station with tunnel, railcar, robot, EOAT, and GPR.
- Develop an inspection sweep with controlled tool orientation and standoff.
- Check collisions, joint limits, singularities, and export a proof video.
Objective 6: Deliver Clear MVP Evidence and Transparency
- Provide system block, railcar concept, EOAT concept, and load case diagrams.
- Document MVP limitations clearly.
- Maintain an assumptions log for track gauge, robot selection, tunnel geometry, and load cases.
Design Prototype Scope Notes
The MVP incorporates a rail-mounted ABB six-axis industrial robot with a custom end-of-arm tool designed to carry a ground-penetrating radar unit. The primary focus of the MVP is the railcar platform: a structural frame positioned on subway rails using steel wheelsets that serves as a stable, load-bearing base for the robot.
The railcar structure is designed to withstand the selected ABB robot’s foundation loads, including forces in the XY plane, vertical force, bending torques in the XY plane, and bending torque about the Z axis under endurance and maximum load cases. Reachability, collision avoidance, joint limit compliance, and repeatable inspection paths are validated in RobotStudio within a constrained tunnel environment. Static structural analysis in Fusion 360 or SolidWorks will evaluate the railcar frame under the worst-case ABB foundation load case to verify strength and stiffness criteria.
Electrical power system design, PLC integration, live sensor feedback, real GPR signal processing, and dynamic vibration modeling are outside the scope of this MVP. The work is limited to validating structural integrity and robotic motion feasibility within a simulated environment.