The Science Behind Modern RTK Corrections

RTK isn’t just about adding more reference stations. It’s about how intelligently you use them to model the atmosphere in between

There are many types of GNSS corrections, such as RTK, PPP-RTK, DGNSS, and more. It can be challenging enough to keep them all straight, but the truth is that even within a single type of correction service there are fundamentally different architectural approaches that affect the accuracy, coverage, and reliability of the service. Understanding these approaches and how they impact the user experience is key to delivering an exceptional location-based product.

The Evolution of RTK: Local → Network → Atmospheric Modeling
Traditional RTK relies on local base stations. While effective, this method is a bit like an analog radio—it works great when you’re close to the transmitter, but the signal degrades quickly as you move away.

Network RTK expands coverage areas by bringing together many local base stations into a single managed network. But to cover a whole country with this approach, you need a "brute force" network of thousands of stations that is incredibly expensive to maintain.

Atmospheric modeling changes the game. Instead of treating every base station as an island, it synthesizes data from an entire network to estimate ionospheric and tropospheric errors over an entire coverage area. With atmospheric modeling, you can achieve the same accuracy and coverage with fewer base stations.

Why Modeling Matters
This approach solves four major industry challenges:

• The 25 km Limit: Corrections from a single base station lose relevance beyond a short baseline. Modeling extends that reach significantly, maintaining precision across distances as high as 70 km.

• The "Jump" Problem: If you’ve ever seen a device’s location "jump" when switching between base stations, you’ve seen a discontinuity error. Modeling enables smooth transitions as devices move between coverage areas, eliminating these erratic shifts.

• Redundancy: With traditional RTK, if a station goes down, corrections are unavailable. In a modeling framework, there is no single point of failure. If one station goes offline, the model automatically adapts using surrounding data to fill the gap.

• Price-Sensitivity: Providers who maintain a massive network of base stations generally pass those costs on to customers. Modeling enables maximum RTK performance with a sparser network of stations, allowing for a more affordable service.

Beyond the Basics
There are two types of atmospheric modeling: geometry-based and physics-based. Both types of models account for how GNSS signals are affected by varying atmospheric conditions, but a physics-based approach is a more sophisticated strategy that considers the composition and physical properties of the atmosphere’s different layers. It takes into account how various atmospheric gases ionize and de-ionize at varying rates under solar radiation and at different altitudes, creating an incredibly robust model that can accurately estimate local atmospheric conditions across an entire coverage area and at any time of day.

In the end, the future of RTK isn’t about building more towers. It’s about building a better model.

And that’s exactly what Swift has done.

Learn more about our atmospheric model: Atmospheric Modeling in GNSS Corrections.

For real-time updates, follow us on LinkedIn

View The GNSS Fix Archive