How accurate is RTK positioning? Explaining errors in 5 points

What practitioners interested in RTK positioning first tend to wonder is, "Can it really measure to the centimeter?" and "How much can it be trusted on site?" To conclude, if conditions are favorable, RTK can achieve high-precision positioning on the order of a few centimeters. However, that accuracy is not constant and is influenced by factors such as satellite visibility, the surrounding environment, correction information, communication status, and the setup method. In other words, precisely because RTK offers high accuracy, it is important to use it with an understanding of where errors originate.

If you leave this point vague and assume "RTK is accurate," you may fail to notice deviations larger than expected, which can affect the quality of as-built verification, staking out, and site condition assessment. On the other hand, if you identify and control the error sources in operation, RTK can be a powerful means of greatly improving on-site productivity. In this article, after organizing guidelines for RTK positioning accuracy, we clearly explain the error factors that should be controlled in practice, divided into five items.

Table of Contents

• What is RTK positioning?

• How accurate is RTK positioning?

• Error 1 Number of satellites and satellite configuration

• Error 2 Sky obstruction and multipath

• Error 3 Distance to Reference Station and Correction Information

• Error 4 Communication status and maintenance of the Fix solution

• Error 5 Installation methods and on-site operation

• Approach to Stabilizing RTK Accuracy in Practice

• Summary

What is RTK positioning?

RTK is a positioning method that aims for higher accuracy than conventional standalone positioning, which determines position using only signals received from satellites. By combining the observation data received by a rover with information from reference stations installed at known points and with network-based correction information, it reduces errors that tend to occur in satellite positioning and obtains high-precision positions in real time. In practice, its use is expanding in tasks that require reliable positioning, such as surveying, stakeout, as‑built management, maintenance management, inspections, and construction records.

The reason RTK is said to be highly accurate is that it does not merely receive satellite radio signals, but also processes finer information called the carrier-phase and reflects the error trends observed at the base station onto the rover. This allows RTK to reduce errors to the centimeter level in situations where standalone positioning can be off by several meters. In practice, people say "switching to RTK suddenly changes usability" because this difference is so large.

However, it should be noted here that the phrase "several-centimeter level" (centimeter-level accuracy (half-inch accuracy)) is not a figure that is guaranteed under all conditions. Very good results can be obtained when satellites are received stably in an open area, correction information is uninterrupted, and the solution is stable. However, under trees, near buildings, in places with many metal structures, or where communications are unstable, the results can differ even with the same RTK. In other words, RTK is not a magic technology; its true performance can only be brought out by understanding the mechanisms of error.

Furthermore, the term "accuracy" as used on site has several meanings. The perspective changes depending on whether you emphasize repeatability—i.e., that repeated measurements of the same point yield similar values—or absolute positional accuracy—i.e., how closely measurements match known points—or whether you evaluate including height. To properly understand RTK positioning accuracy, it's important not to stop at a single phrase like "how many centimeters," but to think through under which conditions, for which metrics, and to what extent errors should be expected.

How accurate is RTK positioning?

If you had to sum up RTK positioning accuracy in one phrase, under good conditions the horizontal position is on the order of a few centimeters, and the vertical direction is a somewhat larger few centimeters as a general guideline. From a practical standpoint, when the sky is open, correction information is stable, and a fixed solution is continuously obtained, the horizontal position becomes fairly reliable. On the other hand, height is more susceptible to satellite geometry and environmental influences, so it needs to be treated more cautiously than the horizontal. This is why, in the field, people often feel that height tends to vary more.

What is important here is not to judge RTK accuracy solely by the numbers listed in catalogs. Catalog values and general performance guidelines are merely benchmarks under specific conditions. On site, satellite geometry, latency in correction data delivery, surrounding reflections, the rover’s setup, and other factors can combine to produce momentary offsets exceeding several centimeters. Especially for tasks where differences of a few centimeters affect decisions—such as stakeout, checks near boundaries, or construction quality control—you should not assume “RTK means it’s okay,” but instead use it while monitoring the solution quality at that time.

Also, RTK accuracy varies greatly depending on status of the positioning solution. A typical example is the difference between a Fix solution and a Float solution. A Fix solution is a state in which integer ambiguity resolution is stable and high accuracy can be expected. In contrast, a Float solution is a state in which the solution has not yet settled and errors tend to be larger. On site, people tend to look only at the coordinates shown on the screen, but in reality checking "what state you are currently measuring in" is the entry point for ensuring accuracy. The fact that numbers are displayed does not mean those numbers are sufficiently trustworthy.

Furthermore, it is safer not to judge RTK accuracy based on a single observation. By taking multiple observations—measuring the same point again after some time, moving slightly and remeasuring, or checking against a known point—you are more likely to uncover cases where numbers that look clean on the surface are actually biased. What is especially important for field practitioners is to understand, not "how many centimeters (inches) under ideal conditions," but "how much variation to expect under their own site conditions." With that perspective, you will neither overtrust RTK nor be unnecessarily afraid of it.

Error 1: Number of Satellites and Satellite Geometry

What you should first grasp in RTK positioning are the number of satellites and their geometry. It’s common to assume that seeing many satellites is reassuring, but what truly matters is not just the count; it’s how they are distributed across the sky. For example, if the visible satellites are biased toward the same direction or you rely mostly on low-elevation satellites, the stability of the position solution will degrade. Conversely, if satellites are well balanced and spread across the sky, the solution quality tends to be higher even with the same number.

Places where this difference tends to appear in practice are mountainous areas, along valleys, near retaining walls or slopes, and locations with densely packed buildings. In environments where only part of the sky is visible, the directions of the satellites that can be received tend to be biased. As a result, even if the planar position is still maintained, vertical accuracy can suddenly deteriorate. When only the height fluctuates unnaturally on site, it is often not just due to setup errors or equipment faults but also related to a deterioration in satellite geometry.

Also, satellite geometry changes over time. What was stable in the morning can suddenly become unstable as the time of day changes. This may not be an equipment problem but rather the result of a change in the combination of satellites in view. In other words, even at the same location, the ease of measurement can vary depending on when you measure. Short tasks are more likely to overlook this, but at survey points where quality is important, simply waiting a little and re-observing can improve the results.

To reduce errors caused by satellite geometry, it is effective not to measure immediately after starting reception but to wait until the solution stabilizes, check quality indicators, and, if necessary, re-observe at a different time. In particular, for important points, avoid fixing them in a single shot; incorporating confirmations separated in time reduces the risk of adopting values from a moment when they happened to look good. The accuracy of RTK positioning is not determined solely by the fact that satellite signals are being received. The sky conditions and the satellite geometry under which observations are made greatly influence how errors manifest.

Error 2: Sky Obstruction and Multipath

The second major source of error is sky blockage and multipath. Sky blockage refers to a condition in which buildings, trees, bridges, slopes, temporary structures, and the like block the sky, making it difficult to receive satellite signals adequately. This is a relatively easy-to-imagine factor, but in practice the more troublesome issue is multipath. Multipath is the phenomenon in which signals from satellites are reflected by surfaces such as walls, metal, vehicles, glass, or puddles, take a longer route, and reach the receiver. The receiver may not be able to fully distinguish between directly received signals and reflected signals, which causes discrepancies in position calculations.

The scary thing about multipath is that it can easily look like reception is working. With simple blockage you’ll notice because the number of satellites drops, but reflections can leave you apparently "receiving" while only increasing the errors. For that reason, on site people tend to assume there’s no problem because numbers are being displayed, and only later notice discrepancies with known points. Be especially cautious near metal fences, steel members, heavy machinery, exterior walls, retaining walls, and other structures. Even if the sky is fairly open, errors increase if there are many reflective objects nearby.

This effect tends to show up in height, but it can also produce a non-negligible bias in the horizontal plane. Moreover, because it manifests differently at each measurement point, a value that was stable at a nearby point can suddenly fluctuate after moving only a few meters. In practice, if you encounter a situation where “only this point doesn’t match for some reason,” you should first suspect nearby reflections. It is not uncommon for the environment at the measurement site, rather than the positioning device itself, to dominate the errors.

As a countermeasure, the basic approach is to observe from a position where the sky appears as open as possible and you can keep a distance from reflective objects. If you absolutely must measure near obstacles, you need to take steps to detect environment-induced bias, such as slightly changing your position and comparing results, varying the measurement time, or observing multiple times. Also, and importantly, even if the numbers look consistent on site, do not decide based on a single measurement; it is safer to reconfirm after changing observation conditions. The first step in reducing errors is to adopt the premise that RTK accuracy is not determined solely by the equipment’s performance but is strongly governed by the radio environment at the measurement location.

Error 3: Distance to the Reference Station and Correction Information

The third source of error is the distance to the reference station and the quality of the correction information. RTK is not a positioning method that can be completed by the rover alone. The reference station also observes the satellites and, based on those observations, applies error corrections to achieve high accuracy. This mechanism is very effective, but as the distance between the rover and the reference station increases, the commonality of the error environment they experience weakens, and the corrections may become less effective. In particular, because the effects of the atmosphere are not spatially completely uniform, the farther the distance, the more likely it is that components that cannot be fully corrected will remain.

In practice, people sometimes become satisfied simply by checking whether correction information is being received, but in reality you need to be aware of the conditions under which that correction information was generated and how appropriate it is for your current work location. Whether you can use a nearby reference station, whether network-based corrections adequately cover the entire area, whether there is any delay in the correction information, and whether there are discrepancies in the coordinate reference settings — these points affect accuracy. These are not random variations but can appear as systematic biases, which are very troublesome in the field.

One case that deserves special attention is when it appears stable to within a few centimeters but does not match the known point. In such cases, even if the observations themselves are stable, there may be problems with the reference-side settings or with how correction information is handled. In other words, high reproducibility and being in the correct position are separate issues. If multiple people at the site obtain the same offset on different days, it is better to suspect the reference frame or correction conditions than individual operator error.

To suppress these errors, it is important to use correction information only after understanding its source and conditions, to verify consistency at known points, and to evaluate especially carefully under long-distance conditions or in mountainous areas. Although RTK is convenient in real time, if you overlook the quality of the correction information, work can proceed while carrying discrepancies that are not easily noticeable visually. If you are going to discuss accuracy in terms of a few centimeters (a few in), you must evaluate not only the receiver but also the very foundation of the corrections.

Error 4 Communication Status and Maintaining a Fixed Solution

The fourth source of error is the communication condition and the maintenance of the Fix solution. In RTK, because high-precision solutions are maintained by continuously receiving correction information, the stability of communication is directly linked to accuracy. If communication becomes unstable, updates to correction data may be delayed or temporarily interrupted, causing the solution state to drop from Fix to Float. Furthermore, even if it appears to recover quickly, it may not be fully stable internally, and adopting the immediately following observations as-is can lead to incorporating errors.

On site, people tend to consider only a complete loss of communication as a problem, but in reality short momentary outages and delays cannot be ignored. Even if correction information is being received, when the update interval is irregular the coordinates can gradually drift or the values can jump at certain times. Moreover, these changes may not appear as prominent warnings on the screen, so if the observer is not accustomed to checking the status display they can be easily missed. To achieve high accuracy with RTK, it is essential to make a habit of looking at the solution status and quality indicators together with the coordinate values.

Also, just because a fix was obtained once does not mean the same quality will be maintained thereafter. While moving, reception can deteriorate, or the device’s orientation or the surrounding environment can change, causing the solution to become unstable. In particular, when observing multiple points in succession while moving, the fact that the previous point had no problem does not guarantee the next one will. You should check the status at each point and, if necessary, wait a little, confirm stability after reinitialization, and reobserve important points.

To reduce this error, it is effective not only to operate in locations with good communication conditions but also to incorporate "state checks" into the observation workflow itself. For example: rather than recording immediately after achieving a Fix solution, wait until it has been stable for a certain period before logging; after moving, do not finalize immediately—pause briefly; at important points, observe whether consecutive values have settled. RTK is a convenient real-time technology, but precisely because it is real time, its quality is affected by the instantaneous state. To preserve accuracy, do not prioritize speed alone; cultivate a practice in the field of continuously verifying that the Fix solution is being maintained.

Error 5 Installation Methods and On-site Operation

The fifth source of error is the installation method and field operations. When people think of RTK they tend to focus on satellites and communications, but in practice human factors are a surprisingly large source of error. For example, a pole being slightly tilted, not being correctly placed directly above the observation point, entering the antenna height incorrectly, confusing the definition of the survey point, or the pole tip sinking into soft ground — any of these lead to deviations of several centimeters (several in). Precisely because RTK is high-precision, these small differences in handling cannot be ignored.

Height in particular is a parameter that is easily affected by installation errors. Even if the horizontal position is reasonably accurate, mistakes in entering the antenna height or a pole that is not plumb can cause the height to shift unnaturally. Moreover, even if the observer believes they are holding it correctly, subtle tilts and swaying are likely to occur on unstable footing or when working in a hurry. At the site, before questioning the performance of the equipment, it is important to first check whether our setup and operation are being carried out with consistent quality.

Also, even with the same equipment and the same location, results will not be consistent if operational rules are ambiguous. One operator records immediately after obtaining a fix, while another waits a few seconds before recording. One operator verifies known points at every station, while another checks them only at the start. When these differences accumulate, the error grows as an operational difference rather than an equipment difference. To use RTK positioning stably, it is essential to implement procedures that ensure the same level of quality no matter who performs the measurements, rather than relying on individual experience.

To accomplish this, it is effective to standardize the entire workflow from pre-observation checks, verification at installation, state checks during recording, to post-observation review. Operational measures such as making re-observation mandatory at critical points, performing start and end checks at known points, cross-checking antenna-height input, and defining re-measurement conditions when abnormal values appear can be carried out as part of routine management rather than as special techniques. RTK errors do not come only from the sky. Ultimately, the thing that determines quality is the operation itself—how measurements are handled in the field.

Practical considerations for stabilizing RTK accuracy

As we have seen, the accuracy of RTK positioning is affected by multiple factors: satellites, the environment, corrections, communications, and setup. Therefore, what matters in practice is not memorizing the "nominal accuracy" of RTK, but being able to judge "how much you can trust it at this particular site, at this particular point, and with this particular procedure." To do that, you should not determine accuracy from a single observation; you need to accumulate and evaluate information that can be verified on site.

The fundamental practice is to verify known points. Confirming alignment at known points before starting work lets you quickly identify any significant issues with that day’s correction conditions, equipment status, or operating procedures. Checking the known points again at the end of the work also makes it easier to determine whether conditions degraded during the job. This may seem like a detour, but it is one of the most effective ways to prevent rework. On sites that are difficult to revisit, completing quality checks on the spot is particularly important.

The next important point is making multiple observations of critical points. For points that serve as positioning references, affect downstream processes, or require accountability, it is safer not to end data acquisition with a single measurement. Taking measurements at different times, returning along a slightly different route, or checking at times with different environmental conditions helps prevent being swayed by a coincidentally good or bad value. RTK is convenient because it processes in real time, but that convenience also means results can be influenced by the conditions at that moment. For that reason, multiple confirmations are effective.

It is also important to change how you use it according to the required accuracy. The level of verification needed differs between tasks where a rough position is sufficient and tasks where a difference of several centimeters (a few in) affects the deliverable. You do not need to treat every point with the same strictness, but it is also risky to treat everything with the same lightness. By clarifying the required accuracy for each task and adjusting the number of observations, verification methods, and acceptance criteria accordingly, you can leverage RTK’s strengths while balancing quality and efficiency.

From the perspective of on-site training, rather than teaching with a single sentence like “RTK is centimeter-level so it’s accurate,” it is important to convey that “centimeter-level (cm level accuracy (half-inch accuracy)) can be aimed for, but it can break down under poor conditions, and there are typical patterns to how it degrades.” If observers know the error factors, they can immediately review the environment and operations when anomalous values appear. Conversely, using the system without understanding the rationale leads to a false sense of security simply because numbers are produced, and quality checks are postponed. To establish RTK in the field, sharing how to interpret accuracy is more important than merely introducing the equipment.

Summary

The accuracy of RTK positioning can reach the centimeter level horizontally under good conditions, and the vertical accuracy can also be high enough for practical use. However, those figures are not automatically obtained at all times. Five factors — the number and geometry of satellites; sky obstructions and multipath; the distance to the reference station and the correction information; the communication status and maintenance of a Fix solution; and the installation method and field operations — combine to determine the actual error. In other words, what governs RTK accuracy is not only the equipment itself, but the observation environment and how it is used.

What matters for practitioners is not simplifying it to “RTK is X centimeters.” It is understanding under what conditions high accuracy can be expected, in which situations errors tend to increase, and how to check observation results, and then embedding that understanding into on‑site procedures. If that is done, RTK becomes not merely a high‑performance positioning method but a practical foundation that supports positioning for layout, verification, recording, and management.

To make RTK practical on-site, it's important not only to understand the theory of accuracy but also to consider portability, ease of integration into daily workflows, and ease of verifying results. If you want to bring high-precision positioning into more familiar, everyday use, it's worth considering options that combine RTK positioning with device operation close to everyday use, such as LRTK (an iPhone-mounted GNSS high-precision positioning device). By correctly understanding error factors and putting verification procedures in place before use, RTK can become a powerful means to raise both the speed and the quality of on-site decision making.