A GNSS receiver showing a fixed RTK solution can transform a day’s work. But what accuracy does RTK deliver when the results need to stand up in a setting-out file, an as-built survey, a volume calculation or an asset record? For most professional field applications, correctly configured RTK can achieve centimetre-level absolute positioning – commonly around 8 mm horizontal and 15 mm vertical, plus a parts-per-million distance component, under favourable conditions.
That specification is highly capable, but it is not a blanket guarantee of 1 cm everywhere on every site. RTK accuracy depends on the correction source, satellite visibility, baseline length, coordinate reference system, antenna setup and field verification. Understanding those dependencies is what turns a capable rover into a dependable survey workflow.
What Accuracy Does RTK Deliver in Practice?
Real-time kinematic positioning uses carrier-phase measurements from GNSS satellites and correction data from a base station or network. The rover compares its observations with the correction source, resolves the integer ambiguities and reports a fixed solution. Once fixed, it can deliver positions far more precise than standalone GNSS.
A realistic expectation for an open-sky site, with a quality receiver and a fixed solution, is approximately 1-2 cm horizontal accuracy and 1.5-3 cm vertical accuracy at the measured point. The vertical component is normally less precise than the horizontal component. This matters when surveying drainage, carriageway levels, earthworks platforms or any feature where reduced level tolerances drive the decision.
Manufacturers often express performance as a figure such as 8 mm + 1 ppm horizontally and 15 mm + 1 ppm vertically. The first figure is the base measurement uncertainty. The ppm element increases with separation from the correction source. At a 10 km baseline, 1 ppm adds 10 mm. Network RTK reduces the practical effect of longer distances by modelling atmospheric errors across multiple reference stations, but conditions and service quality still matter.
It is also essential to separate precision from accuracy. A rover may repeat very closely around the same incorrect coordinate if the wrong coordinate system, transformation or antenna height has been entered. Repeatability is useful, but it is not proof that the data is correctly tied to the required project datum.
The Difference Between a Fixed and Float Solution
The status shown by the controller is not a minor detail. A fixed solution means the receiver has resolved the carrier-phase ambiguities with sufficient confidence. This is the condition required for centimetre-level work. A float solution has not completed that resolution and can be decimetres out, occasionally more.
Do not treat a float coordinate as an RTK survey observation simply because corrections are being received. Wait for fixed status, review the reported horizontal and vertical precision, and re-observe critical points. If the receiver repeatedly drops from fixed to float, the underlying cause needs attention rather than a workaround.
Typical causes include tree canopy, building façades, steelwork, parked plant, overhead obstructions and poor mobile-data coverage. These can block satellites, reflect signals and interrupt correction delivery. A good receiver cannot remove the physical limitations of the sky view.
Why Site Conditions Change the Result
The advertised RTK figure is normally established in controlled, open-sky conditions. Construction and infrastructure environments are rarely that simple. Multipath is one of the most common sources of poor performance: satellite signals reflect from walls, glass, water, metal or machinery before reaching the antenna. The receiver may still report fixed, yet the observation can be biased.
Dense woodland creates another challenge. A rover may maintain a solution under light canopy, but satellite geometry and signal quality can deteriorate rapidly as foliage, branches and terrain obstruct the horizon. In forestry, boundary work or corridor mapping, it is often sensible to combine RTK with total station observations, post-processed GNSS or LiDAR and photogrammetry workflows, depending on the required deliverable.
Satellite geometry also affects confidence. A receiver tracking many satellites does not automatically have an ideal geometry. Satellites concentrated in one area of the sky provide weaker geometric strength than a well-distributed constellation. Checking PDOP, signal quality and solution age helps the operator recognise when a measurement deserves another observation.
Coordinate Systems Matter as Much as GNSS Performance
For UK projects, a centimetre-level measurement in the wrong reference frame is still wrong. RTK corrections may be supplied in ETRS89 or another GNSS reference frame, while project drawings, cadastral information and engineering models may use OSGB36 National Grid. Heights introduce a separate issue: ellipsoidal GNSS height is not the same as orthometric height relative to Ordnance Datum Newlyn.
The controller, correction service and processing software must use a consistent transformation and geoid model. A mismatch can create offsets that are far larger than the rover’s stated RTK precision. These errors can be systematic, meaning every surveyed point appears credible but is displaced by the same amount.
Before starting work, establish the project coordinate system and vertical datum, then confirm how those are configured in the rover. On controlled sites, check known control points before and during the survey. If the project control and RTK observations disagree, do not force the data to fit without identifying whether the cause is a local grid, control issue, transformation setting or antenna-height error.
Antenna Height Is a Small Input With Large Consequences
RTK measures the position of the antenna phase centre, not the ground point, kerb edge or prism point of interest. The controller applies the entered antenna or pole height to calculate the final coordinate. A 20 mm height-entry error produces approximately a 20 mm error in the reported level.
Use a stable pole, a calibrated bubble or tilt-compensation system, and the correct measurement method specified by the equipment. Tilt compensation can improve productivity when capturing detail, but it does not make poor control, multipath or a wrong antenna height disappear. For critical level work, a levelled pole and independent check remain good practice.
Setting Out, Topographic Survey and Machine Control
RTK is particularly effective where the tolerance aligns with centimetre-level positioning and productivity matters. Topographic surveys, utility asset capture, stockpile measurement, agricultural guidance, preliminary site models and many construction setting-out tasks are strong applications. It allows a single operator to capture or stake out large numbers of points quickly without establishing line of sight between instruments.
The correct tolerance should always determine the method. If a task requires a few millimetres, such as precision structural alignment, settlement monitoring or certain rail and industrial installation activities, RTK alone may not be sufficient. A total station, digital level, static GNSS or a combined control approach may be more appropriate.
For machine control, the practical question is not merely the rover’s best-case accuracy. It is the accuracy of the entire chain: design model, site calibration, base or correction network, machine antenna configuration, sensor calibration and regular checks against physical control. A 10 mm rover specification cannot compensate for an outdated design surface or a site calibration with a 30 mm residual.
A Field Procedure That Protects Accuracy
Reliable RTK work is built on routine checks rather than faith in a status icon. At the start of a shift, occupy at least one known point and compare the measured coordinate with the expected value. Repeat the check after moving between work areas, following a correction outage, after changing coordinate settings or when operating close to obstructions.
For important surveyed features, take independent repeat observations. Reoccupation after a short interval, ideally with a different satellite geometry, is a practical way to reveal a questionable result. Record solution status, precision values, correction source, antenna height and any site limitations in the field notes. These details provide a useful audit trail if the data is later queried.
Quality control should continue in the office. Review outliers, check levels against known control and confirm that the exported data uses the specified coordinate reference system. Where the results feed CAD, BIM, GIS or an engineering model, clear naming and coding reduce the risk of a correct coordinate becoming an incorrect deliverable.
Choosing the Right RTK Setup
A local base and rover can give excellent results where the base is established on verified control and the project needs an independent site reference. It also avoids dependence on mobile coverage for local radio-based corrections. However, it requires setup, radio planning and confidence in the base coordinate.
Network RTK is efficient for mobile survey teams working across multiple locations. It removes the need to establish a base each day, provided the correction service is available and the rover has reliable data connectivity. For many UK survey and construction applications, it is the practical default. The decision should be based on control requirements, geography, communications and the consequence of downtime – not only on subscription cost.
LiDAR Tech UK can help assess the appropriate rover, correction method and field-to-office workflow for the accuracy your project actually requires. The most valuable RTK result is not the smallest figure on a specification sheet; it is a verified coordinate that lets the next team make the right decision with confidence.