Say Goodbye to AR Misalignment! Bringing High-Precision AR to the Field with LRTK

Having a smartphone camera that, when pointed at the site, overlays drawings or models onto the actual scenery would be extremely useful. It could greatly reduce the hassle of comparing paper drawings while measuring with a tape or pounding stakes to confirm positions. Such AR (augmented reality) technology is attracting attention in construction and civil engineering sites. However, when actually using AR, users often feel that “the display is slightly off from reality…?” The virtual lines or models you painstakingly displayed can appear displaced by tens of centimeters from where they should be. This AR misalignment problem has been a major obstacle to making AR practical on site.

In recent years, DX (digital transformation) in the construction industry has accelerated through initiatives such as the Ministry of Land, Infrastructure, Transport and Tourism’s *i-Construction*, and efforts to use digital data on site are progressing. AR is expected to contribute to labor saving and intuitive verification on site, but until now its use beyond a “toy level” has been difficult due to positioning accuracy issues. Enter the combination of high-precision GNSS (RTK) and smartphone AR. Smartphones can now achieve positioning accuracy of a few centimeters, dramatically reducing AR display misalignment. This article explains the causes of AR misalignment and its impacts, reviews conventional countermeasures and their limitations, and introduces how high-precision alignment using RTK can solve the problem. Through the latest solution called LRTK, we explain how to bring centimeter-precision AR to the field. Finally, we present examples of applying LRTK to simple surveying.

Table of contents

• What the AR misalignment phenomenon is

• Main causes of AR display misalignment

• Problems caused by AR misalignment

• Conventional countermeasures for AR misalignment and their limits

• Eliminating AR misalignment with high-precision GNSS (RTK)

• Field deployment of centimeter-precision AR with LRTK

• Expanded use cases with LRTK: enabling simple surveying

• Frequently asked questions (FAQ)

What the AR misalignment phenomenon is

When displaying virtual objects in the real world through a smartphone or tablet, the phenomenon of AR displays being offset from the real position can occur. For example, a design line that should be drawn on the ground may appear to float 30 cm (11.8 in) above the actual ground, or a virtual equipment model may appear slightly shifted sideways from its intended installation position. When AR objects do not align perfectly with real objects and gradually drift, the industry often calls this “drift.” If the misalignment is large, the AR display cannot be trusted and becomes unusable for on-site verification or work support.

There are several possible causes of AR misalignment, but most stem from errors in typical smartphone sensors and positioning. The next section looks in detail at the main causes.

Main causes of AR display misalignment

The main reasons current smartphone AR displays become misaligned include the following:

• GPS accuracy limits: The smartphone’s built-in GPS typically has errors on the order of meters. That prevents placing virtual models accurately at designated positions. If the position information is off, the AR display appears at that wrong location.

• Errors in orientation sensors: A smartphone’s electronic compass (magnetic sensor) and gyroscope have errors and drift. Even a small angular error causes larger position offsets for distant objects (for example, if the heading is off by 2°, at 10 m (32.8 ft / 33 ft) ahead the position error can be several tens of centimeters).

• AR mapping errors: AR estimates device movement by capturing feature points in the camera image, but if tracking is not perfect, virtual object positions can gradually shift. Wide movement ranges or feature-poor scenes make mapping unstable and can cause display drift.

• Environmental factors and coordinate inconsistencies: A compass can be affected by surrounding magnetism or steel structures, and mismatches between the coordinate system used in drawings and the field positioning can also cause large offsets. For example, using the wrong reference coordinate in drawings can result in AR displays differing by meters.

As described above, device positioning accuracy, attitude sensor accuracy, and AR engine characteristics inevitably introduce AR display misalignment.

Problems caused by AR misalignment

If AR displays remain misaligned, various issues arise for on-site use. First, trust in the display is lost, so even valuable AR information becomes hard for workers to rely on for site verification. For example, if an area that is actually correctly constructed appears displaced in AR, workers may suspect unnecessary rework, or conversely, actual misalignment may be overlooked due to AR misdisplay—both risky outcomes.

Also, low-accuracy AR is unusable in practice, so even after introduction, sites end up relying on tapes and surveying instruments. If positions are off by meters it’s out of the question, but even tens of centimeters of error cannot be tolerated for construction management requiring precise alignment. In fact, people say, “AR that’s off by meters is unusable in practice, but if errors are within a few centimeters it’s sufficiently reliable for site checks.” In other words, if AR misalignment can be suppressed to the level of several centimeters, it becomes a practical tool for the first time.

Therefore, solving AR misalignment and improving accuracy is the key to fully utilizing AR on site.

Conventional countermeasures for AR misalignment and their limits

Before high-precision GNSS was used, how did sites deal with AR display misalignment? The main conventional measures include the following:

• Sensor calibration: Calibrating the smartphone’s electronic compass before use or resetting gyroscope drift on a level surface. Procedures include waving the device in a figure-eight motion or pointing it at a reference line to reduce sensor errors as much as possible. However, this does not completely eliminate misalignment, and errors reappear after extended movement or environmental changes.

• Alignment using markers or reference points: Placing QR codes or marker boards on site and using the camera to read them to calibrate the virtual model’s position. Alternatively, manually aligning the virtual model to known points on site can correct initial position. This corrects initial offsets but may produce new misalignments when moving over a wide area, and placing markers repeatedly is labor-intensive.

• Combining with simple surveying: Rather than relying solely on AR, some workflows measure key coordinates with a total station or GPS and then offset the AR model to match those results. For example, measuring one point from the drawing on site and entering the difference into the app to correct the whole model. This achieves a certain accuracy, but once surveying is required it can no longer be called “easy AR.”

These measures have some effect but none provide a fundamental solution. Manual adjustments and marker placement are laborious, and device sensors alone cannot maintain stable accuracy over long times or wide areas. They remain at the level of “try to avoid drift,” and are insufficient to completely eliminate AR misalignment.

Eliminating AR misalignment with high-precision GNSS (RTK)

The trump card for fundamentally solving AR misalignment is the use of high-precision GNSS positioning. Among these, the RTK (Real Time Kinematic) method can achieve centimeter-level positioning by correcting satellite positioning errors. While a typical smartphone GPS has meter-level errors, RTK enables extremely high accuracy such as ±1–2 cm (±0.4-0.8 in) horizontally and a few centimeters vertically.

With RTK positioning accuracy, it becomes dramatically easier to align virtual objects to real-world coordinates. When the smartphone knows its absolute position almost perfectly, and you have the coordinate data of the design or model, the AR display can make the virtual and real “match almost perfectly.” What was previously difficult—overlaying drawings onto the real scene with exact alignment—becomes achievable.

Furthermore, high-precision position information suppresses AR engine drift. Continuously feeding absolute coordinates obtained from RTK to the AR system automatically corrects small tracking errors. As a result, users get stable displays where “models do not float in the air even when walking around.” A virtual model placed once will not drift away and require readjustment.

Recently, “smartphone RTK” solutions have emerged that make RTK technology easy to use on smartphones. For example, LRTK is one such solution: a small RTK-capable GNSS receiver that attaches to a smartphone, combined with a dedicated app, enables anyone to perform centimeter-precision positioning easily. The next section looks at how LRTK can introduce high-precision AR on site.

Field deployment of centimeter-precision AR with LRTK

LRTK is a solution that combines an RTK positioning device for smartphones with an AR cloud service to realize AR that does not drift on site. By attaching an LRTK receiver to a smartphone to receive satellite correction data, you can perform RTK positioning on site immediately (if network is available, virtual reference station services can be used; in areas where network is difficult, local base stations or augmentation from Japan’s “Michibiki” satellites can be used). Using the highly accurate current position obtained this way and projecting design data stored in the cloud onto the smartphone AR, models appear at their true-to-scale positions without troublesome pre-alignment work.

For example, if you prepare site boundary line data, LRTK can reproduce that line on the ground almost perfectly through the smartphone. Without drawing temporary lines on the ground, you can immediately check on the smartphone whether there is any discrepancy between the design and the site. Even when walking across a wide area, the model remains fixed to the correct coordinates, so there’s no worry that the edges will be off by tens of centimeters. In short, you can literally “say goodbye to AR misalignment.”

LRTK also considers ease of use on site. The compact device integrates with the smartphone, eliminating the need to carry tripods or heavy equipment. With just a smartphone, site supervisors and engineers themselves can perform positioning and AR display on site, greatly reducing tasks that previously required a surveying team. For example, layout and as-built checks that used to be outsourced can be instantly confirmed with LRTK. This reduces outsourcing waits and scheduling losses, allowing faster PDCA cycles on site.

Thus, by enabling high-precision AR with a single smartphone, LRTK is helping AR technology take root as a practical on-site tool.

Expanded use cases with LRTK: enabling simple surveying

LRTK, which realizes high-precision AR, is not only for overlaying design data but is also expected to be used as a simple surveying tool. A new workflow called smartphone surveying is emerging, allowing surveying and inspection tasks that previously required multiple people to be performed intuitively by a single person.

For example, in pre-construction stakeout, LRTK’s coordinate navigation function displays the distance and direction to the specified coordinate on the smartphone, enabling a single worker to accurately find the stake position. By following on-screen arrows or virtual stakes (AR markers), workers can reach the specified position with centimeter-level accuracy, so even those with limited surveying experience can proceed with confidence. Also, in as-built management, projecting the planned 3D model in AR beforehand lets you immediately check on site whether the finished work matches the drawings. You can instantly see whether embankments or structures have reached the design elevation just by looking at the smartphone. If needed, you can perform point-cloud scanning with LRTK and compare it to the design data in the cloud for quantitative verification on the spot.

LRTK also has a function called “geotagged photos,” which tags high-precision coordinates and camera orientation to photos taken with the smartphone for cloud sharing. Later, when you return to the same place, past photo positions appear as icons in AR, making it easy to identify changes over time or repair locations. These capabilities will significantly transform site management that previously relied on manpower and intuition.

In this way, LRTK is an all-in-one on-site DX tool combining AR visualization and positioning functions. By eliminating the stress of AR “drift” with high-precision AR and enabling seamless surveying, inspection, and recordkeeping, it is expected to dramatically improve site productivity and accuracy. If you’re interested, consider trying the new surveying and construction management experience offered by LRTK.

Frequently asked questions (FAQ)

Q: *What equipment is required for high-precision AR displays?* A: *Basically, you need a smartphone, a high-precision GNSS receiver (RTK-capable device), and a corresponding AR display app. A representative example is an RTK antenna that attaches to a smartphone and a dedicated app, as in LRTK. The smartphone itself does not have to be the latest high-end model, but it should be a device with sufficient performance that supports AR functionality (ARKit or ARCore). Also, for long-duration use, having a mobile battery is recommended.*

Q: *Can anyone use it with just a smartphone? Is special training required?* A: *Compared to traditional surveying instruments, the operation is more intuitive, but some basic training beforehand is advisable. Learning basic app operation, RTK concepts, and operational precautions reduces the risk of confusion on site. That said, the system is designed so non-surveying specialists can operate it. There are cases where site supervisors and construction managers mastered it after a few hours of training and trial use. Initially, it’s recommended to perform trial runs and compare results with experienced personnel, then gradually expand the scope of application for smooth adoption.*

Q: *Do I need to set up an RTK base station every time? Can it be used in sites without network?* A: *It depends on how you obtain RTK correction information. If there are nearby public reference stations or internet access, you can receive VRS (virtual reference station) services that use the Geospatial Information Authority of Japan’s reference stations via the smartphone and achieve centimeter-class accuracy without installing a dedicated base station. In this case, the smartphone must be connected to the internet via cellular or other means. In remote areas with poor connectivity, you can set up a simple local base station (mobile base) and send correction data via radio. In Japan, you can also use CLAS (centimeter-level positioning augmentation service) provided by the quasi-zenith satellite Michibiki to obtain correction data without internet using compatible receivers. LRTK supports multiple correction methods to adapt to site conditions.*

Q: *How closely do AR displays align with reality in practice? I’m worried about errors.* A: *Under favorable conditions, horizontal position errors are on the order of 1–2 cm (±0.4-0.8 in). Vertical accuracy is also within a few centimeters, so in most cases you won’t visually perceive any misalignment. However, this assumes the RTK maintains a solid “Fix” solution and the device’s attitude sensors are properly corrected. Accuracy degrades in environments with poor satellite reception, and compass errors on the device will affect the display. It’s not guaranteed to be perfectly exact at all times, but in normal outdoor environments you can obtain practically sufficient accuracy. The important point is to periodically verify AR accuracy against known points or markers while using it. Even with some error, comparing with on-site references and making corrections as needed allows practical use and increases reliability.*

Q: *Can AR overlay be used in dark places or at night?* A: *Position information is obtained by GNSS, so positioning accuracy does not change at night. However, if the camera image lacks sufficient illumination, AR overlays are hard to see and the device’s AR visual tracking may struggle to maintain accuracy. For night or dark environments, illuminating the target with floodlights or using LiDAR-equipped devices can help maintain stability in some cases. Considering safety, it is preferable to perform surveying and AR checks during daylight when possible. In dark sites, switching to coordinate guidance (numeric navigation) based on high-precision positioning rather than relying on AR imagery is an effective alternative.*

Q: *What about places where GNSS cannot be used, such as indoors or underground?* A: *In environments where satellite positioning is impossible, such as indoors or in tunnels, you cannot obtain absolute positions via RTK. In that case, an alternative is to use AR with local coordinates based on known points. For example, you can set reference points on the floor with a total station, manually input their coordinates into the app, and use that in place of smartphone position to perform basic AR. Recently, indoor positioning technologies using UWB or Visual SLAM for AR are being developed. At present, achieving centimeter-level alignment indoors is not easy, but for some uses planar detection or marker-based AR can be a suitable workaround. The key is to combine other positioning methods where GNSS is unavailable to expand AR applicability.*

Q: *I’m concerned about costs—are the devices expensive?* A: *Compared to traditional surveying instruments or 3D scanners, a smartphone + RTK receiver setup is relatively affordable. Actual prices vary by model and service type, but it often costs a fraction of a dedicated surveying GNSS set. In addition, reducing outsourced surveying, inspections, and rework can lower overall costs. Depending on the site scale, reductions in labor and time may lead to relatively quick return on investment. Some products like LRTK are offered as subscription services, allowing low initial cost trial deployments. We recommend starting with a small-scale site to evaluate effectiveness, then scaling up gradually.*