Experiment Overview
In this experiment, we validated RTK positioning accuracy at the following locations, each featuring different environmental conditions.

• Urban Area (High-Rise District):
  A cityscape surrounded by tall buildings, characterized by significant GNSS multipath from building reflections and frequent signal blockages.

• Mountainous Area (Near Forest):
  A region surrounded by mountains and trees, where the sky is partially obstructed and satellite elevation angles are limited.

• Open Plain (Suburban Vacant Lot):
  A wide, flat area with no tall obstructions in the surroundings—an ideal environment for receiving GNSS signals clearly.

​

In each environment, we performed positioning by comparing the measured coordinates against known control points and evaluated the resulting errors. The true coordinates of these control points were determined in advance with high accuracy using static GNSS surveying against the electronic reference‐station network and total‐station measurements. We then calculated the deviations between those true values and the RTK results obtained with the LRTK. The equipment used was the Lefixea LRTK Pro2, a multi-GNSS, dual-frequency RTK receiver.

The LRTK Pro2 is a compact GNSS receiver with an integrated antenna and battery, supporting network RTK (VRS) and the “Michibiki” Quasi-Zenith Satellite System’s centimeter-level augmentation service (CLAS). Positioning data are logged via Bluetooth to a smartphone app and can be stored and analyzed in the cloud. For this experiment, the LRTK was mounted on a tripod or pole at each measurement location, correction data were retrieved from a network RTK (VRS) service, and real-time positioning was carried out.

Additionally, for comparison, we collected data in standalone (single-receiver) mode without applying RTK corrections to assess the accuracy difference. In both the urban and mountainous sites, measurements were taken at different times of day (e.g., morning and afternoon) to evaluate the effects of satellite geometry and changing surroundings. The positioning results (coordinates) obtained in each environment were compared against the known control-point coordinates, and a statistical error analysis—calculating mean error, standard deviation, and maximum error—was performed.

​

Statistical Analysis of Measurement Results
Our analysis of RTK positioning errors across each measurement environment confirmed that LRTK can determine positions with centimeter-level accuracy in all cases. The error statistics by environment are shown below (errors are calculated separately for the horizontal plane and the vertical direction).

​

Measurement Environment Mean Error (H / V) Standard Deviation (H / V) Maximum Error (H / V)

Urban Area (High-Rise District) 5.2 cm / 7.8 cm 3.1 cm / 5.4 cm 12.4 cm / 18.9 cm

Mountainous Area (Forest) 3.8 cm / 6.5 cm 2.7 cm / 4.1 cm 9.6 cm / 15.3 cm

Open Plain (Unobstructed Area) 2.1 cm / 3.5 cm 1.4 cm / 2.0 cm 5.0 cm / 7.1 cm

​

Environmental‐specific RTK positioning error statistics (using LRTK, deviation from known points): In urban areas, multipath from tall buildings led to larger errors than in other environments, occasionally exceeding 10 cm. Conversely, in open plains, errors were very small, with both horizontal and vertical mean errors contained within just a few centimeters. Although errors increased slightly in mountainous areas, mean errors still remained at the centimeter level. These results show that under favorable conditions, RTK‐GNSS can achieve accuracy on the order of a few centimeters, and even in environments with some satellite signal constraints, it can maintain practically usable accuracy (errors of several to a dozen centimeters).

​

Moreover, when compared with standalone positioning (uncorrected GNSS fixes), the accuracy improvement achieved by RTK is immediately apparent. For example, in the open plain, standalone positioning showed an average horizontal offset of about 0.8 m, with maximum errors approaching 1.5 m. In contrast, RTK positioning kept errors to an average of 2–3 cm, dramatically narrowing the error range. In urban areas, standalone fixes also exhibited deviations of several meters, but once RTK was applied, errors were reduced to a few centimeters to a few tens of centimeters. In other words, the data clearly demonstrate that the relative positioning corrections provided by RTK greatly enhance GNSS positioning accuracy.

​

Accuracy Comparison Between RTK and PPP
In addition to RTK, another high‐precision positioning method is PPP (Precise Point Positioning). PPP enhances standalone GNSS accuracy by using global orbit and clock error models (augmentation data broadcast by satellites), and has the advantage of not requiring a base station. However, its drawback is that convergence to centimeter‐level accuracy takes time—typically several minutes to tens of minutes in real time.

In our experiment using the LRTK Pro2, we also tested positioning with the QZSS “Michibiki” centimeter‐class augmentation service (CLAS, a PPP-RTK approach). We found that after about five minutes of static observation, horizontal accuracy reached within 10 cm and vertical accuracy within 15 cm. However, it did not achieve the few‐centimeter precision in a short time that RTK delivers. These results align with the general limitations of PPP, where achieving sub‐10-cm accuracy typically requires several minutes to tens of minutes of observation.

​

On the other hand, network‐RTK initializes within a few seconds and achieves centimeter‐level accuracy, making RTK ideal for dynamic positioning applications (such as moving‐object tracking or immediate surveying tasks). The positioning error of RTK‐GNSS (VRS method) commonly used in field surveys is consistently around 3–4 cm, and our LRTK experiment confirmed comparable accuracy. Therefore, in an RTK vs. PPP comparison, RTK clearly wins in terms of immediacy and precision stability (especially for dynamic surveying and real‐time control applications). Of course, when it’s difficult to establish base stations over a wide area, PPP or PPP‐RTK remains effective—so choosing the right method for the application is crucial.

​

Discussion and Conclusions
Through this experiment, we confirmed that RTK positioning with the LRTK device achieves measured accuracies of just a few centimeters in both the horizontal and vertical axes. In open environments, positioning errors were so small that the difference between design-drawing coordinates and on-site measurements was negligible. Even under challenging conditions—such as in urban high-rise districts or forested areas—errors stayed within a range of only a few to a dozen centimeters, fully satisfying the precision requirements of standard construction and surveying operations.

This performance can be attributed not only to the LRTK receiver’s capabilities—such as dual-frequency GNSS and multi-constellation support—but also to the effective use of network RTK corrections and CLAS augmentation.

However, we also identified considerations for maintaining positioning stability. For example, in urban areas, stepping into a building’s shadow sometimes interrupted the fixed solution and degraded RTK accuracy. To stabilize positioning in such cases, it’s important to optimize the antenna’s placement and mounting method. Installing the antenna in a location with an open view of the sky and placing a simple metal‐plate ground plane beneath it can reduce multipath effects. Additionally, enabling multi-GNSS (not just GPS but also GLONASS, Galileo, QZSS “Michibiki,” etc.) helps ensure a sufficient number of satellites are always in view.

Modern receivers like LRTK support multiple GNSS constellations, helping to prevent accuracy degradation by increasing the number of tracked satellites even in urban canyons or valleys. Additionally, operational measures—such as briefly reinitializing the positioning solution at a known control point to reset the fix, and using a leveling bubble or tilt-compensation on the rover pole to maintain verticality—are also effective for preserving accuracy.

As demonstrated above, the LRTK’s measured accuracy and practicality are exceptionally high, proving more than capable for standard surveying and construction applications. To utilize RTK positioning reliably in the field, you need appropriate device settings and operational considerations—but once those are in place, it offers dramatically greater efficiency and labor savings compared to traditional survey methods.

Looking ahead, further advancements in satellite positioning services—such as an increased number of satellites and the adoption of new correction techniques—are expected to enhance RTK’s convenience even more. For example, in Japan from 2024 onward, the QZSS “Michibiki” constellation will expand to seven satellites, enabling more stable centimeter-level positioning nationwide. Riding this wave of technological progress, high-precision GNSS devices like LRTK will play an increasingly vital role in infrastructure maintenance and construction digitalization.

​

Utilization and Implementation Benefits of LRTK
With LRTK’s reliability proven by this accuracy validation, deploying the system can deliver the following advantages for surveying and data‐collection operations.

• Streamlined, Labor-Saving Surveying: Tasks that traditionally required two people—such as control-point surveys or setting out batter boards—can now be performed solo with an LRTK device and a smartphone. For example, on a construction site, a worker simply walks while holding a pole or wearing a helmet outfitted with the LRTK receiver, capturing high-precision coordinates in rapid succession. This reduces manpower requirements while also enhancing safety.

• Immediate Data Sharing: The location data captured with the LRTK app are uploaded to the cloud, so there’s no need to carry field measurements back to the office each time. You can review results on the spot and, if anything is unclear, take additional measurements immediately—enabling a highly flexible workflow. During infrastructure inspections, photos taken on site are automatically tagged with centimeter-precision coordinates and shared to the cloud, allowing managers to accurately assess field conditions from the office.

• Multi-Purpose Use Cases: LRTK is not only transforming construction and surveying workflows but is also poised to make a significant impact in railway and highway infrastructure maintenance. For example, railway operators routinely measure track deformation and settlement; with LRTK, technicians can simply walk alongside the rails to record track displacement with high precision. In highway management, teams can patrol pavement cracks and asset locations, capturing positioning data and notes that directly inform repair planning. Beyond these applications, LRTK can be paired with UAV drones for aerial surveying or integrated with AR solutions for subsurface utility detection—enabling entirely new centimeter-level positioning–based services.

• Ease of adoption: LRTK is designed for use by personnel without specialized surveying training. Its intuitive smartphone interface allows operators to control the system and visually confirm results on a map. A free trial and comprehensive support network are available, ensuring that even companies adopting RTK technology for the first time can get up and running with confidence.

​

It can be used not only on survey sites but also across a wide range of applications, such as automatically tagging infrastructure inspection photos with precise location data. Our accuracy validation has shown that LRTK is a highly reliable, centimeter-level positioning solution in real-world field conditions. By making RTK technology—once expensive and difficult to use—easily accessible, LRTK is expected to accelerate digital transformation (DX) in both the construction industry and infrastructure maintenance sectors.