Estimated reading time: 7 minutes
If your phone says you're 'here' but your job needs a point that's correct within a few centimeters, standard GPS isn't good enough.
What is RTK
Real-Time Kinematic (RTK) is a satellite positioning technique that enhances the precision of GNSS (GPS, GLONASS, Galileo, BeiDou) signals to provide, in real-time, centimeter-level accuracy, typically achieving 1–3 cm horizontally and 3–4 cm vertically. It uses a stationary base station to send correction data to a mobile rover, compensating for atmospheric delays and orbital errors. Normal GPS/GNSS positions are often 5–10 meters off because they rely mostly on code-phase measurements and uncorrected atmospheric and satellite errors.
Key Aspects of RTK
- High Accuracy: Unlike standard GPS, which has a 5-10 meter error, RTK reduces positioning error to the centimeter level.
- System Components: Requires two receivers—a stationary base station at a known position and a moving rover—working together.
- Fix is the goal: Single → Float → Fix; Fix enables the best repeatability and cm-level relative accuracy in good conditions.
- Real-Time Correction: Corrections are calculated and transmitted instantly (often via cellular network or radio) to the rover.
- Applications: Essential for surveying, construction, autonomous vehicles, and precision agriculture to enable precise mapping, navigation, and guidance.
- RTK vs PPK: RTK provides corrections in real-time for immediate navigation, whereas Post-Processed Kinematic (PPK) applies corrections after the data is collected.
RTK Meaning
Normal GPS/GNSS positions are often 5–10 meters off because they rely mostly on code-phase measurements and uncorrected atmospheric and satellite errors. RTK positioning fixes this by using carrier-phase measurements plus real-time corrections from a nearby reference station or network. RTK stands for (Real-Time Kinematic) is a relative positioning technique where a rover receiver uses carrier-phase GNSS signals and correction data from reference station(s) at known coordinates to compute centimeter-level positions in real time. People say "rtk real time kinematic gps," but modern RTK is multi-constellation GNSS (GPS, GLONASS, Galileo, BeiDou) and often multi-frequency (L1/L2/L5), which improves time-to-fix and fix stability. Anyone who's done layout or mapping in the real world knows the pain points: staking a curb line—float solutions cause offsets that show up as rework; drone mapping—RTK reduces ground control needs, but only if corrections are reliable; robotics—latency and fix drops can cause path-tracking errors. Before you choose a base, a network, or an NTRIP subscription, here's the practical 60-second model of how RTK works in the field.
RTK Technology
Most buying mistakes happen because teams think RTK is a "feature you turn on." In practice, RTK performance comes from a system: receiver capability, correction delivery, and the environment you're working in.
RTK Real Time Kinematic explained in 60 seconds
An RTK system is (1) a reference station at a known coordinate, (2) a rover that measures satellites, and (3) a correction link that delivers near-real-time observation data so the rover can resolve carrier-phase ambiguities.
- Measure: Base/network measures the same satellites as the rover.
- Compute: Base/network computes correction information (observation-based, not just a single offset).
- Stream: Corrections are sent as RTCM 3 over NTRIP (internet) or radio.
- Resolve: Rover runs ambiguity resolution; solution status becomes Single → Float → Fix.
- Work: In Fix, horizontal is typically centimeter-level in good conditions.
Example you'll recognize from the field: a survey rover connects to an NTRIP caster on a job site and reaches Fix in ~5–30 seconds in open sky. Why it matters: Fix depends on corrections + satellite geometry + environment, not just "having RTK hardware." If your sky view is bad, your correction age is high, or your baseline is too long, the receiver can't do magic.
Single RTK, RTK Float, RTK Fix in the real world
- Single RTK: autonomous GNSS (meters).
- RTK Float: carrier-phase used but ambiguities are not integers yet (often decimeter to sub-meter).
- RTK Fix: integer ambiguities resolved; best repeatability and cm-level relative accuracy.
Operationally, train crews to log two things: solution status and correction age. Don't stake critical points in Float, and don't accept a "Fix" that's paired with stale corrections or constant re-initialization. Also: don't claim "1 cm guaranteed", just say "cm-level in Fix under good conditions."
Fix depends on corrections, satellite geometry, and the environment—not just "having RTK hardware."
RTK GPS vs Standard GNSS At-Glance
- Accuracy: standard GNSS = meters; DGNSS = sub-meter; RTK = cm-level (Fix).
- Infrastructure: RTK needs a correction stream + an RTK-capable receiver (often multi-frequency, multi-constellation).
- Time behavior: RTK requires initialization; it can drop in tough environments (trees, cranes, urban canyon).
Quick workflow comparisons (this is the "which tool when" view):
- Standard GNSS: simple, but typically meter-level.
- DGNSS: code-based corrections; improves to sub-meter but doesn't solve carrier-phase integers.
- RTK: real-time carrier-phase with ambiguity fixing for cm-level relative results.
- PPK: post-processed kinematic; uses carrier-phase but computes after the mission.
- PPP: precise point positioning; uses precise orbits/clocks and converges more slowly, often without a local base.
That last point is key for planning: if you can't maintain a correction link (remote spraying, offshore, long UAV corridors), you may need a workflow that tolerates outages. Professional RTK services like RTKdata.com provide 20,000+ reference stations across 140+ countries, helping you find correction coverage near your job sites without guessing.
What RTK corrects (and what it can't)
RTK works because many GNSS errors are shared over short distances (the baseline):
- Satellite clock bias and orbit/ephemeris: similar for nearby receivers.
- Ionosphere and troposphere: delays are correlated over short distances, so differencing reduces them.
But some problems are local and won't disappear just because you're in Fix:
- Multipath: reflections near the antenna are local; a base station can't remove them for the rover.
- Poor geometry (high PDOP): fewer satellites or bad angles reduce solution quality even with corrections.
Symptom mapping you can trust: open sky + low multipath → fast Fix; under trees/urban canyon → frequent Float/Fix drops and longer initialization. That's also why rtk real time kinematic positioning accuracy in cm is realistic only when the environment lets the receiver keep clean phase tracking.
Why Multi-Band Helps Fix
Multi-frequency (L1/L2/L5) allows the receiver to model ionospheric delay better and improves ambiguity resolution speed. Multi-constellation increases satellite count and geometry, reducing PDOP and improving Fix reliability. There's a point of diminishing returns. After dual-band + multi-constellation, antenna quality (phase center stability), ground plane, mounting, and the local multipath environment often matter more than adding another constellation.
RTCM, NTRIP, Latency Setup
Once you've chosen your workflow, performance usually comes down to one thing: do you have clean satellite signals and a stable correction stream? This is where RTCM, NTRIP, latency, and mountpoint selection matter.
RTCM corrections in practice
RTCM is a standard message format for GNSS correction data; RTCM 3 (often written as rtcm3) is the modern version used for RTK. If you're evaluating rtcm corrections for rtk real time kinematic, you're really evaluating whether your rover can ingest the message set your correction source outputs.
- 1005/1006: reference station coordinates (base position messages).
- MSM 107x/108x/109x/112x: multi-signal observation messages for GPS/GLONASS/Galileo/BeiDou (MSM messages).
- 1230: GLONASS code-phase bias info that helps interoperability.
Most crews don't pick messages manually. They confirm rover and mountpoint compatibility, then watch bandwidth and correction age so the rover doesn't starve or lag.
NTRIP basics: caster, mountpoint, client
NTRIP is a protocol for streaming GNSS correction data over the internet. It's the most common answer to how to get rtk corrections via ntrip for survey, construction, ag, UAV, and robotics.
- NTRIP caster: the server that hosts streams.
- Mountpoint: a named correction stream you connect to.
- Client: your rover/controller software that logs in and receives RTCM.
- Enter host/port and your username/password.
- Select a mountpoint that matches your rover capabilities (constellations, MSM, etc.).
- Confirm you're receiving RTCM at a steady rate and your correction age stays low.
- Enable NMEA GGA output if your caster uses it for VRS positioning.
Common misconception: line-of-sight between base and rover is not needed with NTRIP; you only need cellular/internet plus good satellite sky view. Trees and buildings can wreck GNSS even if your LTE signal is perfect.
Transport options and latency impact
- Internet/cellular: usually easiest; watch correction age (seconds) and dropouts.
- UHF radio / LoRa: good where cellular is weak; requires radio setup and often line-of-sight for the radio link.
- L-band corrections: broadcast option for wider areas; useful when internet is unreliable (implementation depends on receiver/service).
RTK assumes corrections match the rover's measurement time. High latency or jitter increases time-to-fix and can cause Fix to drop during motion. Drone and robotics example: a drone moving at speed with 2–3 seconds of correction delay can see unstable heading/position; stable low-latency corrections improve path tracking. For rtk real time kinematic for drones mapping, this is the difference between clean camera event geotags and a dataset you end up "saving" with extra GCPs.
Picking The Right Mountpoint
If offered, start with a mountpoint that matches your receiver's constellations (GPS+Galileo+BeiDou, etc.) and uses MSM where possible. If you're near a specific station, a nearest-station stream can work well; if you move across a region, VRS may stay more consistent. If you're trying to figure out how to find the best rtk correction service provider by country, don't start with marketing claims—start with practical coverage: station proximity, density, redundancy, mountpoint types, and expected latency on your local cellular network. For implementation specifics and terminology, reference https://docs.rtkdata.com/.
Conclusion
RTK real time kinematic positioning achieves cm-level results by using carrier-phase measurements plus real-time corrections and resolving integer ambiguities (Fix). Your day-to-day performance depends on correction quality (latency/uptime), baseline distance, satellite geometry (PDOP), and local multipath. Absolute coordinates depend on correct base coordinates and the right coordinate datum/geoid; many "bad RTK" reports are configuration mismatches. Next step: find a correction source near your job sites and validate Fix stability—RTKdata.com makes it easy to identify available stations and RTK networks worldwide (20,000+ stations, 140+ countries), then you can use the checklists above to confirm performance in the field.
Frequently Asked Questions
What is the meaning of RTK in surveying?
Real-time kinematic positioning (RTK) is a satellite-based positioning technique used in surveying to correct for common errors in GNSS systems. RTK delivers centimeter-level accuracy in real time by using correction data from a base station, making it essential for precise land surveying, construction layout, and geodetic control work.
What is the most frequent mistake made in GPS control surveying?
Multipath error is the most frequent mistake in GPS control surveying. Multipath errors occur when GNSS signals reflect off surfaces such as buildings, vehicles, water, or the ground before reaching the receiver. These reflections lead to skewed positioning data, particularly in urban or built-up environments, significantly degrading GNSS accuracy.
What is RTK (Real-Time Kinematic) positioning?
Real-Time Kinematic (RTK) is a satellite positioning technique that enhances GNSS systems to provide centimeter-level accuracy, compared to standard GPS, which has meter-level errors. It works by using a stationary base station to transmit corrections to a moving receiver (rover) in real time.
How does RTK achieve centimeter-level accuracy compared to normal GPS?
Normal GPS relies mainly on code-phase pseudorange (meter-level noise). RTK uses carrier phase and resolves integer ambiguity; when ambiguities are fixed, the rover can use the carrier's short wavelength precision. A nearby reference station helps cancel shared errors (satellite clock/orbit and atmospheric delays), which is why RTK workflows outperform autonomous GNSS.
What is the difference between RTK and DGPS/DGNSS?
DGNSS (DGPS) applies code-based corrections and typically reaches sub-meter accuracy. RTK systems use carrier-phase observations and ambiguity resolution for cm-level results in Fix. RTK is more sensitive to signal quality and needs consistent correction delivery.
What do RTK Float and RTK Fix mean?
RTK Float means the rover is using carrier phase but ambiguities aren't locked as integers; accuracy is often decimeter to sub-meter. RTK Fix means the receiver has resolved integer ambiguities (ambiguity resolution), enabling centimeter-level relative accuracy. Fix can drop to Float due to obstructions, multipath, correction latency/outages, or cycle slips.
How far can a rover be from an RTK base station?
Baseline distance is the rover-to-reference separation; as it grows, atmospheric differences increase and fixing gets harder. A common practical range for strong performance is 10–20 km from a single base in typical conditions, with longer possible but less robust. Network RTK reduces distance sensitivity by using multiple stations and modeling.
Do I need line-of-sight between base and rover for RTK?
Only if you're using a radio link (UHF/LoRa) between base and rover; radios often need line-of-sight for reliable transmission. With NTRIP corrections over the internet, you do not need line-of-sight to the base station—what matters is GNSS sky visibility and data connectivity. Trees and buildings affect satellite signals even if your internet link is fine.
What is RTCM and what RTK correction messages matter?
RTCM is the standard format for GNSS corrections; RTCM3 is commonly used for RTK. Typical messages include 1005/1006 (reference station coordinates), MSM 107x/108x/109x/112x (multi-signal observations), and 1230 (GLONASS bias info). Users mainly need rover/mountpoint compatibility and stable corrections rather than memorizing message numbers.
What is NTRIP and how do RTK corrections get to my receiver?
NTRIP (Networked Transport of RTCM via Internet Protocol) streams RTCM corrections over IP. Your rover/controller is an NTRIP client that logs into an NTRIP caster and selects a mountpoint; many VRS streams require the rover to send NMEA GGA positions upstream. Low latency and low jitter improve time-to-fix and fix stability.
What is the difference between Network RTK and using my own base station?
Your own base gives control and can work without cellular if you use radio, but you must set correct base coordinates and manage setup. Network RTK offers broader coverage, redundancy, and less setup, but depends on service availability and data connectivity. Choose based on where you work (single site vs region), baseline needs, and reliability requirements.
RTK vs PPK: Which should I use for drone mapping?
RTK applies corrections in real time and can geotag imagery precisely during flight if connectivity is solid. PPK processes after the flight; it's often better when cellular is unreliable or you want to reprocess with different settings. Many teams fly RTK when possible and keep logs for PPK as a backup—especially when aiming for tight deliverables.
Why is my RTK accuracy bad even when it says Fix?
Multipath (reflections) can bias measurements even in Fix—move the antenna away from metal, vehicles, walls, and tree canopy edges. Datum/geoid mismatch can make heights or coordinates look "wrong" even if the solution is consistent; verify coordinate datum and geoid model. Wrong antenna height/ARP or a bad base coordinate will shift every point consistently.
Can RTK work under trees or in cities with tall buildings?
It can work intermittently, but canopy and urban canyon environments reduce satellite visibility and increase multipath, causing Float solutions and Fix drops. Multi-constellation and multi-frequency help, but they can't fully overcome heavy blockage. For robotics, consider sensor fusion (IMU/odometry) to ride through RTK outages.