The Problem With Standard GPS

Your smartphone’s GPS is accurate to within about 5 metres on a good day. That is fine for turn-by-turn driving directions, but useless for a robot that needs to follow a specific path through a garden, navigate a row of grapevines, or land precisely on a charging pad. For those tasks, you need a different approach: Real-Time Kinematic GPS, or RTK-GPS.

RTK-GPS achieves accuracy down to 1-2 centimetres. That is roughly 250 times more precise than your phone’s GPS, and it is the technology that has unlocked an entire generation of wire-free robots — most notably wire-free robotic mowers like the Mammotion Luba 2 AWD 5000.

What Makes Regular GPS Inaccurate

GPS works by measuring how long radio signals take to travel from satellites 20,000 km overhead to your receiver. Tiny errors in that measurement translate to position errors:

  • The atmosphere slows the radio signals by an unpredictable amount.
  • The satellite clocks are slightly off.
  • The receiver itself has timing noise.
  • Signals bouncing off buildings (multipath) create false readings.

These errors accumulate to roughly 5 metres in good conditions and much worse in cities or under tree cover.

The RTK Trick

RTK’s insight is brilliant: if you place a second GPS receiver at a known fixed location nearby (called the “base station”), it experiences the SAME atmospheric and satellite errors as your robot. The base station can calculate exactly how wrong its GPS reading is at every moment, then transmit that correction to the robot.

The robot subtracts the same correction from its own GPS reading and ends up with centimetre-accurate position. The closer the robot is to the base station, the better the correction works — most consumer systems are accurate within 5-10 km of the base.

What You Need to Run RTK

A consumer RTK setup typically includes:

  • The robot with a multi-band RTK-capable GPS antenna (different from standard GPS).
  • A base station usually mounted on a tripod or rooftop with a clear sky view.
  • A communication link between them — usually radio, sometimes Wi-Fi or 4G.
  • Multi-constellation support — modern RTK uses GPS, GLONASS, Galileo, and BeiDou simultaneously to maximise satellite coverage.

For larger commercial deployments, networks of “virtual reference stations” provide RTK corrections over cellular networks — meaning the robot does not need its own base station, just a SIM card.

Where RTK Excels

  • Robotic lawn mowing: Wire-free mowers like the Worx Landroid Vision L1600 can follow precise mowing patterns instead of random bounces.
  • Agricultural robotics: Tractors can plough rows within 2 cm of each other, eliminating overlap and saving fuel.
  • Drone surveying: Precise geo-tagging of aerial photos for mapping and inspection.
  • Delivery robots: Curb-side navigation that knows exactly where the sidewalk edge is.

The Catch

RTK requires clear sky view. Under dense tree canopy, in narrow urban canyons, or indoors, RTK signal quality degrades sharply. Modern systems fall back to vision-based positioning or inertial dead-reckoning when RTK is unavailable — but those backups are not as accurate.

This is why robotic mowers struggle under heavy tree cover and why drone manufacturers specify “RTK-fix percentage” alongside flight time.

Why It Matters Now

Until recently, RTK base stations cost $1,000-$5,000. The current generation of consumer RTK base stations from Mammotion, Husqvarna, and others costs $300-$600 and works out of the box. That price collapse is what made wire-free robotic mowers viable as a consumer category.

Expect RTK to keep spreading — into delivery robots, smart agriculture, and eventually even autonomous home equipment we have not thought of yet.