GNSS site positioning: what a centimetre is worth

A GNSS rover on a pole used to be a surveyor’s specialist tool, carried by one trained person who set out the site at the start of a job and was rarely seen again until the next milestone needed checking. Today the same underlying technology sits on excavators, pavers and setting-out crews across the site, and it changes how a site is measured from the first day of groundworks to the last day of finishes.

How the accuracy is achieved

Standard GPS on a phone is accurate to a few metres, useful for navigation but nowhere near good enough for construction. Survey-grade GNSS achieves centimetre accuracy by comparing the satellite signal received on site against a correction signal from a fixed base station or a network of reference stations, a technique generally called RTK, real-time kinematic positioning. That correction is what turns a signal accurate to metres into one accurate to a couple of centimetres, and it is also why GNSS accuracy depends entirely on maintaining a good connection to that correction source, not just on a clear view of the sky.

Where it pays off

Setting out on open ground is the clearest win. One person with a rover can place points that once needed two people and a total station, one holding the prism and one operating the instrument, cutting setting-out labour roughly in half on a job with a lot of open-ground points to place. The same coordinates that drive setting out can feed directly into machine control systems, so that an excavator’s bucket or a paver’s screed knows exactly where the design surface is in real time, without a machine operator having to work from staked-out points or grade stakes at all.

Where it does not

Satellite signals do not reach inside buildings, and they struggle noticeably beside tall structures, where reflections off glass and steel facades, sometimes called multipath interference, can degrade accuracy without an obvious warning sign to the person using the rover. Once the frame of a building is up, the crew still needs a total station or a laser for accurate work on the floors themselves, and continuing to trust GNSS in that environment without cross-checking against a known point is one of the more common ways positioning errors creep into a project.

What good practice actually looks like

Tie every measurement on the site back to a single, documented control network, established once at the start of the project and never redefined informally by a crew that finds it inconvenient to reference. Check the rover against a known, previously surveyed point at the start of every shift, not just when a problem is suspected, since a small systematic error introduced overnight, whether from equipment, correction service outage or a configuration change, is far cheaper to catch in a thirty-second check than to discover after a day of work has been set out from a wrong reference.

Agree explicitly who owns the coordinate system before the first stake goes in: which datum, which local grid, which benchmark the whole site’s measurements are tied to. Confusion between coordinate systems, more than any equipment fault, is the most common root cause of a setting-out error large enough to actually matter on site.

The bigger picture

Most positioning errors on site are not caused by the satellites, the correction service or the equipment; they come from someone using the wrong reference point, an outdated control network, or a rover that was never checked at the start of the shift. Investing in the discipline around GNSS use, rather than just the equipment itself, is what actually delivers the accuracy the technology is capable of.

This article is general information and not surveying advice.