Your Phone’s GPS Could Soon Navigate Like a Self-Driving Car

If you have ever watched your blue dot jump erratically across a map while walking through downtown, bouncing from one side of the street to the other like a confused tourist, you are experiencing what researchers call an “urban canyon.” Now, scientists at the Norwegian University of Science and Technology have figured out how to make cheap GPS receivers as accurate as equipment that costs thousands of dollars more.

The implications go beyond fixing annoying navigation glitches. This technology could make autonomous vehicles viable in cities, where current GPS systems struggle with the glass and concrete jungle that surrounds them.

When Buildings Attack Your Signal

The problem is straightforward but vexing. GPS signals travel at the speed of light in a straight line, but tall buildings turn city streets into pinball machines for radio waves. The signals bounce off windows and walls, taking longer routes to reach your phone. Since GPS calculates position based on how long signals take to arrive, these detours throw everything off.

Cities are brutal for satellite navigation.

That is how Ardeshir Mohamadi, the doctoral fellow leading the research, describes the challenge. Traditional GPS can be accurate to within a few meters in open areas. But drop into a narrow street flanked by skyscrapers, and that precision evaporates. For a human with eyes, this is merely annoying. For a self-driving car making split-second decisions, it could be catastrophic.

The Norwegian team tested their system in the streets of Trondheim, comparing cheap consumer-grade receivers (the kind you might find in a fitness watch) against professional equipment costing tens of thousands of dollars. They put both types through two urban scenarios: standard city conditions and what they termed “high-interference urban canyons,” where narrow passages between tall buildings created the worst possible environment for satellite signals.

Throwing Away the Noisy Parts

The researchers’ solution involves a counterintuitive approach: ignoring most of the GPS signal. Traditional GPS uses a code embedded in the radio transmission to calculate distance. But that code is precisely what gets scrambled by reflections. So the team’s system throws it out entirely and focuses only on the carrier wave itself, the underlying radio frequency that carries the code.

Think of it like trying to hear someone across a noisy room. Instead of straining to catch every word (the code), you focus on the rhythm and pitch of their voice (the carrier wave). You lose some information, but what you do hear is clearer.

This “phase-only positioning” technique has been around for years, but it required either wide-open spaces or expensive professional equipment. The Norwegian researchers figured out how to make it work with mass-market receivers in the toughest urban environments by combining it with other correction methods.

One key piece came from an unexpected source. While the Trondheim team was refining their approach, Google quietly launched a service using 3D building models from nearly 4,000 cities worldwide. These models predict how satellite signals will bounce around urban environments. The researchers incorporated this data into their system alongside corrections from Europe’s Galileo satellite network and their own algorithms.

The results surprised even the researchers. In standard urban conditions, cheap receivers using their SmartNav system achieved 100 percent reliability for centimeter-accurate positioning. Even in the most challenging environments tested, where traditional GPS methods delivered accurate positions only 3 percent of the time, the new approach boosted that to 58 percent for consumer receivers and 78 percent for professional gear.

PPP-RTK reduces the need for dense networks of local base stations and expensive subscriptions, enabling cheap, large-scale implementation on mass-market receivers.

That statement from Mohamadi points to the broader significance. Previous solutions for precise urban GPS required networks of ground stations and costly subscription services. This approach works with free satellite corrections and data that tech companies are already collecting for other purposes.

The technology is not perfect. In the most difficult test environment, consumer receivers still failed to lock onto an accurate position 42 percent of the time. And the system needs a clear view of at least four satellites, which is not always possible on narrow streets. But the improvement over standard GPS is substantial enough that it could finally make autonomous vehicles practical in dense urban areas.

The research also revealed an interesting quirk about GPS hardware. When comparing cheap and expensive receivers in challenging conditions, the expensive equipment did not always perform better. Both struggled with severe interference, suggesting that in truly difficult environments, smart software matters more than premium hardware. This finding could accelerate adoption, since it means existing smartphones and consumer devices could potentially support the technology with software updates.

Whether this ends up in your pocket depends on manufacturers adopting the approach. The researchers published their SmartNav positioning engine as a research tool rather than a commercial product. But with GPS-dependent technologies proliferating, from delivery drones to augmented reality applications, the pressure to solve urban navigation problems is only growing. The next time your map shows you walking through a building instead of past it, remember: the fix might already exist. It just needs someone to ship it.

Journal of Spatial Science: 10.1080/14498596.2025.2536567