What Is the Actual Accuracy of RTK-GPS? 8 Key Points to Understand Errors and How to Choose by Application

For practitioners considering implementing RTK-GPS, the biggest concern is probably "how accurate will it actually be?" Even if the specifications look highly accurate on paper, results in the field can vary greatly depending on whether the sky is unobstructed, whether there are obstructions nearby, whether communications and correction information are stable, and how well operators can follow the correct procedures. In other words, RTK-GPS accuracy cannot be judged by a single number.

In this article, we organize how to assess RTK-GPS accuracy in practical work, combining application-specific error sense with perspectives on how to choose a system. People who enter the search keyword "RTK GPS 精度" are likely looking for judgment criteria they can use in the field rather than theory. Therefore, to prevent post-deployment failures such as "it doesn't achieve the expected accuracy," "the height doesn't match," or "there are concerns about repeatability," we explain eight key points you should grasp from a practical viewpoint.

Table of Contents

• First, understand the real-world accuracy of RTK-GPS

• Item 1 Determine allowable tolerances for each application

• Item 2 Consider horizontal accuracy and vertical accuracy separately

• Item 3: Focus on reproducibility rather than single measurements

• Item 4: Understand how errors due to the reception environment manifest

• Item 5 Verify the stability of correction information

• Item 6 Verify initialization time and recoverability

• Item 7 Select operability suited to on-site work

• Item 8: Consider the ease of recording and sharing

• Summary

First, Understand the Real-World Performance of RTK-GPS Accuracy

RTK-GPS is a method that combines correction information with position information obtained from satellites to achieve higher accuracy than standalone positioning. From a practical standpoint, under favorable site conditions when corrections are stable, reception is good, and the solution is reliably fixed, centimeter-level positioning on the horizontal plane can be expected (cm level accuracy, half-inch accuracy). On the other hand, the vertical (height) direction tends to be less favorable than the horizontal, and in actual operation it is necessary to treat vertical measurements with an extra degree of caution compared with horizontal ones.

However, the important point here is not to take the phrase "a few centimeters" at face value. The accuracy of RTK-GPS is affected by many factors, including sky visibility, surrounding structures, trees, vehicles, slopes, movement of machinery, communication stability, the state of correction information, how measurement points are taken, how the pole is held, and checks for consistency with known points. In other words, even with the same equipment the results will change when the site changes, and even at the same site the results can fluctuate depending on the time of day or work procedures.

Also, in practical work, it is not enough to know “how many centimeters the displayed coordinates are off.” What is required is whether that position is actually usable. For example, the required accuracy varies depending on whether the position is being kept only as a current-condition record, used for construction layout, used for verification of completed work, or linked to a maintenance management ledger. The important thing is to judge RTK-GPS accuracy not as an absolute value but by its suitability for the intended use.

Furthermore, accuracy has two aspects: the values that appear accurate on-site and whether you can return to the same position when you re-measure later. If you introduce it based only on the former, unexpected discrepancies will occur in the field. Therefore, to evaluate the true capability of RTK-GPS, it is essential to make a comprehensive judgment from multiple perspectives: intended use, horizontal and vertical aspects, repeatability, reception environment, correction information, recoverability, operability, and recordability.

Item 1: Decide allowable tolerances for each application

The first thing you should do when choosing an RTK-GPS is clarify what you will use it for and how much error you can tolerate. If you start comparing options while this is still unclear, you may impose unnecessarily high requirements or, conversely, choose a configuration that lacks sufficient accuracy for practical work.

For example, if the main objective is to be able to trace locations later—such as recording site rounds, gaining a rough understanding of equipment positions, or linking positions to photographic records—it does not necessarily need to be accurate to within a few centimeters. Even with some offset, it can still be sufficiently useful in practice. Conversely, for tasks such as verifying construction positions, managing locations close to boundaries, checking as-built conditions, or work that takes earth volume or elevation differences into account, even small errors can affect decisions.

What's important here is not to assume that "because RTK-GPS is high-precision, it can do everything." On site, the phrase "several-centimeter class" can have different meanings. It may be sufficient for applications that fix position in the horizontal plane, but insufficient for applications that require strict control of elevation. Or, while measuring a single point may be fine, for applications that continuously collect multiple points to examine alignment or slope, point-to-point variations can no longer be ignored.

Therefore, before making a selection, it is important to first concretize the intended use within your company or on-site. Simply clarifying which tasks it will be used for, whether horizontal positioning is important, whether height is important, whether a single measurement will suffice, whether the same point will be handled repeatedly, and whether there are many obstructions around will significantly change how you assess the required performance.

A common pitfall for practitioners is advancing implementation discussions centered on equipment. In reality it should be the opposite: first decide the intended use, set the allowable error for that use, and then choose an operation that can meet those conditions. When judging the accuracy of RTK-GPS, the starting point is not "how many centimeters it can achieve" but "whether that error is acceptable for this task."

Item 2 Consider horizontal accuracy and vertical accuracy separately

When considering the accuracy of RTK-GPS, many people try to understand it as a single number. However, in practical work you must consider horizontal accuracy and vertical accuracy separately. If you operate without this perspective, you'll encounter the typical problem of "the position appears correct but only the height is wrong."

Generally, RTK-GPS tends to be more stable in the horizontal direction, while the vertical (height) direction is more susceptible to changing conditions. This is because height is affected not only by satellite geometry but also by various factors such as pole tilt, instrument height and the handling of input values, terrain conditions, assumptions in coordinate transformations, and the way survey points are set up. On site, planar positions may look fine while the height slowly drifts and is only noticed during later processing.

Especially in operations where control of elevation differences and slopes is critical—such as earthworks, paving, drainage, slopes, site development, and around foundations—this difference cannot be ignored. A difference of a few centimeters (a few inches) in height alone can affect the direction of flow, the appearance of the finished work, and construction decisions. On the other hand, for tasks whose primary purpose is determining horizontal position—such as routine inspections and location recording—the requirements for vertical accuracy may be relatively lower.

Therefore, when making comparisons, rather than using an abstract description like "this instrument is high-precision," you need to separate how you treat the horizontal direction from how you treat the vertical (height) direction. In practice, it is effective to verify the horizontal (plan) and the height (vertical) separately at known points or check points, and confirm which is stable to what extent. Especially when height is emphasized, it is essential not to judge based on a single measurement result but to take multiple checks over time.

Those responsible for successfully implementing RTK-GPS understand this distinction early on. Rather than treating the word “accuracy” as a single label, separating the evaluation of horizontal (planar) usability and vertical (height) usability leads to greater acceptance in the field.

Item 3 Assess reproducibility rather than single measurements

What is often overlooked when comparing RTK-GPS units is judging their performance based on a single measurement. What really matters in the field is not that you get a good value once, but reproducibility — that you return to similar results when measurements are taken at different times, when remeasured, or when handled by a different operator.

For example, even if the coordinates at a given survey point look clean after a single measurement, if that was simply due to coincidentally favorable conditions, discrepancies may arise during remeasurement on a different day or when reentering from a different route. What becomes a problem in practice is these inconsistencies at remeasurement. Whether it is staking out, as-built verification, ledger updates, or maintenance record-keeping, it is important that the previous and current measurements connect seamlessly.

To check reproducibility, it is effective to operate by measuring the same point after a short delay, returning at a different time to measure, or deliberately moving once and then remeasuring. From the results, whether the measurements consistently fall into roughly the same pattern each time or exhibit large deviations only under specific conditions will show whether the configuration is usable in practice. It is important to look not only at single best values but also at the range of variation.

Reproducibility is also affected by operator-dependent factors. Even with the same equipment, results can differ depending on whether the pole is set up correctly, whether sufficient time is allowed for verification, whether measurement points are chosen appropriately, and whether the operator can judge how to avoid obstructions. In other words, reproducibility is not only a measure of equipment performance but also an indicator of the completeness of operational design and training.

Before deployment, what you should really look at is not “how accurate it was at its best” but “how close a typical operator can get to the same result, each time, in typical field conditions.” If you want to know the true capability of RTK-GPS accuracy, it is essential to evaluate it by the practical standard of reproducibility rather than by a one-off impressive result.

Item 4 Know how errors occur due to the reception environment

RTK-GPS accuracy varies greatly depending on the reception environment. If you don't understand this, you may mistakenly conclude that the equipment's performance is poor, or conversely assume that there's nothing wrong with how it's being used. In reality, sky openness and the surrounding reflective conditions strongly affect how errors manifest.

The most stable conditions are in environments with a wide open sky, few tall buildings or trees nearby, and few metal surfaces or walls that can act as reflectors. In such locations, if correction information is also stable, RTK‑GPS is more likely to reach its full potential. Conversely, beside buildings, under trees, near bridges, below slopes, in valley‑shaped terrain, in areas densely packed with mechanical equipment, or in places with frequent vehicle traffic, satellite visibility can become biased and signals are more likely to be affected by reflected waves.

What's particularly troublesome in the field are cases where, although it appears that the sky is visible, in reality only one direction is open. In such cases, the configuration of satellites that can be received becomes biased, and even if the display looks plausible, the coordinates can gradually become unstable. Also, the vertical direction is easily affected by these influences, so you can have a situation where the horizontal plane appears correct but the height alone is hard to trust.

Therefore, on sites using RTK-GPS, the placement of survey points themselves is important. If you must measure close to an obstruction, it can be effective to establish a stable position a short distance away and then apply the necessary offset from there. Also, rather than measuring immediately after leaving an obstructed area, simply waiting for the solution to stabilize can improve the results.

What matters for practitioners is not eliminating errors entirely, but recognizing environments that are prone to errors and knowing how to handle them. If your assessment of the reception environment is inadequate, no matter how high-performance a configuration you deploy, it will not inspire confidence on site. You need to understand that RTK-GPS accuracy is determined not only by equipment selection but also by how you interpret site conditions.

Item 5 Confirm the Stability of Correction Information

Correction information is essential for RTK-GPS to achieve high accuracy. No matter how good the receiver’s performance is, if the acquisition of correction information is unstable, the expected accuracy cannot be maintained. Therefore, when discussing accuracy, you must evaluate not only the unit itself but also the stability of the correction information.

There are several approaches to managing correction information, such as receiving wide-area corrections via communication lines or using known position information around the site as a reference. What matters here is not the name of the method but whether it can be used stably under your site conditions. In mountainous areas, on reclaimed or developed land, around underground structures, in regions with weak communications, or at sites where work often involves movement, acquiring and maintaining corrections is directly linked to accuracy.

When correction information is unstable, problems arise such as the solution being difficult to stabilize, the state degrading midway, frequent reinitializations, and wait times occurring at each measurement. These not only slow down work but also reduce confidence in the measurement results. On site, operators may hurriedly take measurements while the system state is still inadequate, which later manifests as errors.

It is not enough to only check whether corrections are connected. What matters is whether it remains stable after a circuit of the site, whether it can recover easily after movement or occlusion, and whether it can be used under similar conditions in the same area on different days. Configurations with unstable communication or corrections may seem deployable at first, but in busy field environments they tend to be used less and less.

From a practical perspective, when comparing the accuracy of RTK-GPS you should assess the stability of the entire operation—including correction information—rather than the capabilities of the equipment alone. High precision is achieved only when the receiver, corrections, environment, and procedures are all in place. If the stability of the correction information is neglected, dissatisfaction with accuracy will almost always surface in the field.

Item 6 Verify initialization time and recoverability

The practical performance of RTK-GPS depends heavily on the initialization time required to reach a fixed solution and how quickly it can recover after signal blockage or communication outages. This is often overlooked, but it is an important factor that affects ease of use in the field.

For example, in setups that require long waits every time at morning startup or when beginning measurements after moving, operators will gradually start to avoid that equipment. Moreover, if the system’s status degrades after passing under trees or near structures and you have to wait for it to recover each time, the time required per point becomes unpredictable. As a result, this causes significant stress on sites with many measurement points or where a single person wants to work at a steady, fast pace.

The reason recoverability is important is that the worksite is not always under ideal conditions. It would be fine if tasks were confined to open areas, but in reality movement, obstacles, temporary structures, vehicles, slope shoulders, cut-and-fill, stored materials, and the like are all mixed in. Since conditions can be disrupted each time, the ability to recover quickly is as important as accuracy itself.

Also, initialization time and recoverability aren't simply matters of being fast. If it seems fast but isn't stable, you'll incur more work later for re-measurement and corrections. What matters is that it reaches a stable state in a short time and that you can have confidence in the subsequent measurement results. In practical work, there are many situations where waiting a little and obtaining reliable data is more valuable than apparent speed.

When deciding whether to adopt a system, you should verify it under movements as close to actual operation as possible. Rather than measuring just once in an open area, you should evaluate it in flows such as moving with obstructions in between, visiting multiple points in succession, and returning after some time; otherwise you won't understand the differences you'll face on site. If you want to leverage RTK-GPS accuracy in practical work, you need to regard initialization time and recoverability not merely as performance metrics but as elements directly tied to productivity.

Item 7 Choose operational suitability for on-site work

One aspect that can make a surprising difference when introducing RTK-GPS is operational usability. No matter how good the theoretical accuracy, if it is difficult to use in the field, personnel will become unduly cautious; as a result, the number of measured points will decrease, verification will become lax, and the range of applications will not expand. Whether it becomes established in practice is determined not only by accuracy but by whether it can blend into daily work without causing extra burden.

For example, whether it will be carried by a single person, whether tasks are divided among several people, whether you want to capture many points in a short time, or whether you plan to operate it together with photographic or inspection records will change the required ease of use. On site, the flow—from morning preparation, travel, selection of measurement points, measurement, verification, saving, to sharing—is continuous, so if there is a lot of hassle along the way, the operation will not be sustained regardless of accuracy.

Moreover, ease of operation is directly linked to measurement quality. If changing settings is complicated or verifying the status is difficult, operators tend to proceed with the feeling that “it should have been captured.” As a result, when reviewed later it may become apparent that the solution’s state or the measurement conditions were insufficient. Conversely, a configuration that makes status checks and decisions about re-measurement easy also helps ensure accuracy on site.

Furthermore, it is important to consider whether it is easy to check consistency with known control points, whether it can accommodate site-specific coordinate conditions, and whether procedures remain consistent even when handled by multiple people. RTK-GPS is not a tool that functions solely for positioning; it is a tool that has meaning within the site’s operational rules. Therefore, it is essential to assess not only the level of accuracy but also whether the operations make that accuracy easy to reproduce.

From the standpoint of a field practitioner, "high accuracy under ideal conditions" is less valuable than "the regular operator can use it reliably and effortlessly at their usual site." When selecting RTK-GPS, you shouldn't focus only on desk-based performance comparisons; you should also judge whether it matches the site's workflow and pace of operations.

Item 8 Consider even the ease of recording and sharing

The benefits of introducing RTK-GPS are not determined solely by the moment of measurement. Practical value only emerges when you include how the acquired coordinates and elevations are recorded, reviewed, shared, and linked to accountability. Even with high-precision positioning, if records are vague and cannot be traced later, it cannot be fully utilized on site.

For example, if there is no record of who measured which measurement point, at what time, and under what conditions, it becomes difficult to trace the cause when discrepancies are found later. If photos are not linked to positions, the history of re-measurements is unknown, or coordinate conditions are not communicated when sharing, then even data collected with high precision will not become a practical asset.

Especially in construction and maintenance management, being able to explain the results to others is more important than the measurement itself. The question is whether the values recorded on site can be kept in a form that in-house staff, partners, managers, and successors can understand. In other words, the true capability of RTK-GPS is not just positioning accuracy, but an overall strength that includes the transparency of records and the ease of sharing them.

The ease of recording also affects the number of measurements taken in the field. If record-keeping is cumbersome, workers will record only the bare minimum. Conversely, if positioning results, photos, notes, point names, timestamps, and so on are naturally linked in the workflow, an audit trail will accumulate on site without extra effort. That accumulation makes later verification, quality control, and reporting easier.

When choosing an RTK-GPS, it's important not to look only at positioning performance, but to consider whether the information you want to record on-site can be recorded there without undue effort. A configuration that is truly easy for field personnel to use is not only highly accurate, but also allows the results to be naturally integrated into their workflow.

Summary

You cannot simply sum up the true performance of RTK-GPS accuracy with a single phrase like "how many centimeters (inches) it will deliver." Required tolerances vary by application, and the perspective differs between horizontal and vertical directions. Furthermore, beyond single measurements, verifying repeatability, understanding how errors arise from the reception environment, ensuring correction information is stable, that initialization and recovery are robust enough for practical work, that the operation fits the site's workflow, and that recording and sharing can be managed smoothly — all of these affect the actual "usable accuracy."

In short, when choosing an RTK-GPS the important thing is not chasing numbers that look good under ideal conditions. It is assessing whether, in everyday fieldwork, you can repeatedly deliver the required accuracy for the necessary situations using the required procedures. With this perspective, failures such as "it doesn't achieve the expected accuracy," "height measurements are unstable," or "it stopped being used on site" become much easier to prevent.

If you want to make the whole workflow on site—from position checking to recording and sharing—more practical, it’s important to consider not only positioning accuracy but also usability and how easily the data can be leveraged. From that perspective, iPhone-mounted GNSS high-precision positioning devices like LRTK are also a compelling option to consider. For field personnel who want to make location information more accessible in their day-to-day on-site work, and to go beyond merely measuring to linking measurements with recording and sharing, these devices are an easy-to-consider means of turning RTK-GPS accuracy into on-site value.