Smartphone RTK × AR for Collision Limit Management DX: Changing Railway and Road Infrastructure Inspections

What the collision limit is: definition and operational significance in railways and roads

In infrastructure management for railways and roads, the term "collision limit" refers to the space (clearance) that must be kept free to allow trains or vehicles to pass safely. In the railway field, a cross-sectional space slightly larger than the maximum exterior profile of running vehicles (vehicle gauge) is established, and no structures or obstacles must exist within this area. The safety space surrounding the vehicle is the building clearance, also called the collision limit. For example, a common standard is approximately 2 m (6.6 ft) to the left and right from the track center and approximately 6 m (19.7 ft) in height in electrified sections (4.5 m (14.8 ft) in non-electrified sections) or more; tunnels, viaducts, platforms, and catenary poles are all installed outside the building clearance.

The same applies to roads: the Road Act and related regulations define certain spaces above and beside roads as building clearance. Typically, obstacles must not be installed lower than 4.5 m (14.8 ft) above the road surface or lower than 2.5 m (8.2 ft) above sidewalks. These are the minimum clearances to guarantee the safe passage of large vehicles and pedestrians, and the spaces beneath bridges, viaducts, traffic signals, signs, and tree branches are managed so they do not encroach into this space.

Managing the collision limit (building clearance) is extremely important from the perspectives of safe operation and accident prevention. On railways where vehicles travel at high speed, a shortfall of a few centimeters can lead to contact incidents. Similarly on roads, low overpasses or sagging wires and branches can collide with large trucks. In the past, when new vehicles were introduced, previously unnoticed clearance shortages in existing equipment were discovered and emergency relocation work was carried out. As these cases show, ensuring and regularly checking the clearance margin (safety buffer) is a critical task directly linked to infrastructure safety and security.

Conventional collision limit inspection methods and their challenges

On site at railways and roads, analog methods have mainly been used to check collision limits. In railways, specialized measuring tools, rulers, or a frame called a limit gauge have been used to measure the distance from the track to structures one by one. For example, crews press a specified gauge against the track and visually check gaps with tunnel walls, catenary poles, or platform edges, or they measure the distance from the rail center to wall surfaces with tape measures. On high-speed railways and large networks, dedicated building clearance measurement vehicles (gauge cars) have been run; these use movable rods (feathers) attached to the vehicle body to detect obstructions by contact. On roads, it has been common to measure heights to bridge girders and traffic signals using elevated work platforms or height rods, or to read clearances from on-site design drawings (paper plans) and verify them in the field.

However, many issues have been pointed out with these conventional methods. First is the problem of workload and time. On railways, personnel must enter tracks between trains or at night to take measurements, and manually checking long sections requires enormous labor. Even when using gauge cars, operating costs are high and use is limited. On roads, high-elevation measurements require traffic control, making rapid complete checks difficult. Second is the problem of accuracy and comprehensiveness. Manual work cannot avoid human error, and measurement points are limited, creating risks of oversights and measurement mistakes. Methods that rely on paper drawings at the site may not fully capture construction errors or shifts due to aging. In fact, there have been reports where equipment judged acceptable on drawings was actually below the specified clearance in reality and nearly contacted vehicles. Third, regarding safety, measuring on tracks or roadways is itself dangerous for workers. Rushed measurements in limited nighttime work windows increase the risk of human error, and conventional methods have their limits.

Advantages of visualization using point cloud data and AR and examples of use

In recent years, advances in ICT have significantly changed methods for checking collision limits. A key technology is visualization using point cloud data (three-dimensional survey data) and AR (augmented reality). Point cloud data are collections of many points acquired by LiDAR (laser scanners) or photogrammetry, representing structures and surrounding environments as high-density 3D models. In railways, attempts are increasing to mount laser sensors on dedicated measurement vehicles or trolleys running along tracks to scan tunnel inner walls, catenary poles, and station facilities to millimeter-level shapes. In the road sector, mobile mapping systems (MMS) mounted on vehicles can capture the entire road space as point clouds, including the undersides of viaducts and tree overhangs.

Using point cloud data makes it possible to accurately identify on a desktop the locations that intrude into the collision limit, which previously could only be determined on site. By virtually overlaying the building clearance cross-sectional frame on dedicated software and comparing it with the point cloud, it becomes apparent which parts infringe the standard clearance. For example, applying the prescribed building clearance cross-section to scanned tunnel data allows automatic measurement of vertical and lateral clearances at each section, extracting the most critical locations with minimal margin. For roadside objects such as trees or signs under bridges, actual clearance from the road surface can be calculated from the point cloud, and points below 4.5 m (14.8 ft) can be flagged with warnings.

AR display overlays such point cloud data and clearance lines onto the real world. By viewing the site through the screen of a tablet or smartphone and superimposing virtual guides or models, information that was previously invisible becomes intuitively visible. For example, in railway maintenance, looking at the trackside scenery through a smartphone might show the outline of the building clearance glowing on the screen, indicating the positional relationship with equipment in real time. Such AR can clarify even a few centimeters of margin that are hard to perceive with the naked eye. In practice, there are reports of loading tunnel shape data acquired as point clouds into AR-capable tablets to visually check the margin at the tunnel wall on site with a transparent overlay. This allows small protrusions that were hard to discern on paper cross-sections to be visually detected on the spot, aiding priority decisions for repairs.

Practical RTK positioning with smartphone + LRTK and instant distance checks

To realize AR visualization accurately in the field, it is necessary to determine device position and orientation with high precision. Recently, on-site implementation of RTK positioning combining smartphones and high-precision GNSS has attracted attention. RTK (Real Time Kinematic) is a technology that corrects satellite positioning (e.g., GPS) errors through simultaneous observation with a base station to achieve centimeter-level (half-inch accuracy) positioning in real time. Previously, fixed survey instruments were required, but now smartphone-linked compact GNSS receivers have emerged, and smartphone RTK solutions (such as LRTK) make high-precision positioning easily available to anyone.

Using smartphone + LRTK positioning dramatically streamlines on-site instant distance checks. For example, to check the distance between a station platform edge and the track center, conventional methods required a tape measure or total station. With smartphone RTK, holding up a smartphone on the platform can provide current coordinates, and horizontal distance to a pre-set track centerline can be calculated instantly. The screen can display information such as "○○ cm (○○ in) to the track" or "○○ cm (○○ in) remaining to the building clearance," allowing measurement and recording simultaneously.

The same applies to road height checks. Under viaducts or in tunnels, using a smartphone + LRTK and pointing the phone at the ceiling allows the device to calculate elevation and the relationship to the underside of the girder and display "clearance ○○ m (○○ ft)" in real time. Without carrying dedicated equipment, field engineers can check clearances at multiple points in a short time with just a smartphone, helping prevent inspection omissions. Moreover, RTK positioning data from LRTK can be integrated with the smartphone's AR functionality, so virtual guides (clearance lines and distance displays) shown in AR align with real-world positions with high accuracy. This makes the positional relationship between equipment and clearance lines seen through AR reliable and greatly improves the accuracy of on-site decisions.

Procedure for detecting collision limit intrusion risks and interference judgment from point clouds

Now let's look at the specific procedure for performing interference judgment (detection of insufficient clearance points) from point cloud data. Broadly, the steps proceed as follows.

• Point cloud data acquisition: For railways, scan along the track with track inspection vehicles or LiDAR-equipped drones; for roads, use MMS vehicles or ground-based lasers to acquire point clouds of the road space. When necessary, measure the entire target area densely and create 3D models of the track/road and structures.

• Coordinate alignment and reference frame setup: Align the acquired point cloud to a known coordinate system (survey coordinates or design drawing coordinates). At the same time, define on the computer the cross-sectional shape that constitutes the collision limit (for railways, the building clearance cross-section; for roads, a horizontal plane at 4.5 m (14.8 ft) above the roadway, etc.) as a reference frame. In railways, arrange cross-section frames continuously along the running direction, and on roads, set a horizontal plane at a fixed height along the road surface.

• Extraction of intrusion points: For each point in the point cloud, calculate the relative position to the reference frame. For railway cross-sections, check whether point cloud points fall inside the frame at each section; for roads, search for points whose height from the road surface is below the specified value. Dedicated software can automate this judgment and mark locations where clearance falls below thresholds.

• Risk level assessment: For each extracted intrusion point, measure how much it infringes the building clearance (how many cm it exceeds). Priorities differ depending on whether a location is barely within limits or clearly exceeding the standard, so classify risk levels with numerical evaluations or color coding.

• Verification and sharing of results: Technicians verify extracted results and, if necessary, perform on-site re-inspection or root-cause analysis. Even though the judgment is virtual on the point cloud, it may be desirable to visually confirm the actual structure in some cases. After verification, problem locations are shared among stakeholders and measures such as repair work or tree trimming are planned promptly.

These steps allow comprehensive identification of intrusion risks into collision limits that were previously overlooked. In particular, on railways it is necessary to consider the cant effect in curves, where the vehicle tilts and the vehicle body expands outward; point cloud analysis software can incorporate such dynamic expansions of the vehicle gauge into simulations. On roads, it is necessary to allow certain height margins considering the heights and load shifts of large vehicles, but detailed height distributions from point clouds enable safety-oriented assessments and appropriate countermeasures.

On-site AR display for verifying safety margins and sharing awareness among stakeholders

Clearance information derived from point cloud analysis and on-site measured data can be further utilized via AR display. On-site AR display overlays analysis results and guidelines onto live images on a tablet or smartphone, enabling all stakeholders to verify safety margins from the same viewpoint and share a common understanding.

For example, consider an inspection under a railway viaduct. A field technician launches an AR app on a smartphone and points it at the girder underside. The screen displays pre-analyzed building clearance lines and shows color-coded comparisons indicating how much vertical clearance exists relative to the actual girder underside. If some locations fall below allowable heights, those parts are highlighted in red with annotations such as "this is ○○ cm (○○ in) below the standard." Multiple staff members viewing this footage together can immediately share where the problem is and how many centimeters of margin are lacking. This visual approach is easier to understand than verbal explanations or drawings and helps prevent miscommunication due to differing perceptions.

AR is also an effective communication tool during meetings with related departments for projects such as station platform renovation. If construction managers and designers point a tablet at the platform and confirm the positional relationship between the platform and the vehicle gauge line together, they can reach a shared understanding of hazardous areas and where additional safety measures are needed. AR screens can be recorded as photos or videos and later shared in meeting materials. Such visualization reduces information gaps between field and office, enabling faster and more appropriate decision-making regarding collision limit management.

Cloud integration of inspection results and streamlining records and reports

Cloud integration is also an important element in DX (digital transformation) of collision limit inspections. If point cloud data, on-site positioning data, and AR confirmation results are centrally managed in the cloud, recordkeeping and reporting tasks can be greatly streamlined. Traditionally, measurement records handwritten on site had to be brought back and recompiled into reports. Now, clearance values measured with a smartphone or tablet are saved digitally on the spot and automatically sent to a cloud database. If photos and AR screen screenshots are stored with tagged location and timestamps, it becomes possible to accurately trace "when, where, and what the problem was" afterward.

Furthermore, sharing data on a cloud platform smooths reporting to internal and external stakeholders. For railway operators, headquarters technical managers can check inspection data and remaining clearances at each point from the cloud without visiting the site. Alerts mark detected anomalies, and photos and point cloud viewers allow detailed examination so prompt responses can be coordinated. In municipal road management, inspection contractors' in-field inputs can be shared in real time within the municipal office, enabling immediate orders to remove high-priority obstructions.

The digitalization and centralization of data also benefit lifecycle management. Comparing with past inspection histories makes it easy to analyze trends such as "that tunnel's clearance has decreased further" or "tree overhangs are increasing year by year." This supports planning preventive maintenance earlier, or conversely reducing inspection frequency in locations showing no issues, enabling risk-based efficient maintenance. Cloud integration gets the PDCA cycle for collision limit management running digitally, allowing technicians who used to be tied up with report writing to spend more time on higher-value analysis and planning.

Toward introduction of collision limit management DX with smartphone RTK and AR

As shown above, new methods utilizing smartphone RTK, AR, and point cloud analysis are poised to bring innovation to collision limit management for railway and road infrastructure. Transitioning from conventional manpower-intensive inspections to smart inspections based on digital data and modern technologies enables both safety and efficiency. Solutions like LRTK, which combine smartphones with compact GNSS receivers, are groundbreaking in that they enable centimeter-level surveying and AR visualization on site without specialized equipment. Because collision limit management DX can be started immediately using readily available smartphones without heavy investment, it is accessible even to regional railway operators and local governments.

Detecting obstructions within collision limits and measuring clearances is a highly responsible task where mistakes can lead to serious accidents. For that reason, DX that visualizes conditions, prevents missed measurements, and enables data-driven decisions is required. Starting with familiar technologies such as smartphone RTK + AR allows field teams to immediately realize reduced burdens and improved accuracy. Expanding step by step to include point cloud surveying and cloud systems can lead to a comprehensive digital transformation of infrastructure inspections. Collision limit management DX is a key initiative to raise railway and road safety to the next level. Amid concerns about aging skilled workers and labor shortages, digital technologies make it possible for anyone to perform safety checks at a uniform level, and they are effective for skill succession. This is a good opportunity to incorporate smartphone RTK and AR into the field and take a step toward achieving infrastructure maintenance that provides peace of mind for everyone.