How GPS Works: The Satellites, Signals, and Math Behind Location Accuracy

GPS is one of those technologies so embedded in daily life that it’s become invisible. Turn-by-turn navigation, package tracking, emergency services dispatch, precision agriculture, financial transaction timestamping, and air traffic control all depend on it. Your phone can tell you exactly where you are on Earth, accurate to a few meters, from signals sent by satellites in orbit 20,000 kilometers away. The fact that this works at all—let alone that it works continuously, globally, and for free—is a remarkable engineering achievement worth understanding.

GPS is not magic, and it’s not complicated in principle. The underlying idea is elegant and the mathematics are straightforward. The engineering required to make it work reliably and accurately at planetary scale is where the complexity lies.

The Fundamental Principle: Trilateration

GPS is based on trilateration—determining a position by measuring distances from known reference points. If you know you’re 5 kilometers from a tower, you could be anywhere on a circle (in 2D) or sphere (in 3D) of radius 5 km centered on that tower. If you also know you’re 4 km from a second tower at a different location, you’re at one of two intersection points. A third measurement from a third reference point typically narrows it to a single location. A fourth measurement is used to resolve the timing issue (explained below).

GPS satellites serve as the reference points. Each satellite continuously broadcasts a radio signal containing two pieces of information: who is transmitting (which satellite) and when the signal was sent (a precise timestamp). Your GPS receiver picks up these signals, notes when each one arrived, and calculates how long the signal took to travel from each satellite. Since radio signals travel at the speed of light (~300,000 km/second), travel time × speed of light = distance. With accurate distances to four or more satellites whose positions are known, the receiver can calculate your position.

The Constellation: How Many Satellites and Why

The US GPS system (officially called NAVSTAR GPS) operates a constellation of 31 active satellites in medium Earth orbit at an altitude of approximately 20,200 km. The satellites are arranged in six orbital planes, each inclined 55° relative to the equator, distributed so that from anywhere on Earth’s surface, between 6 and 12 satellites are typically visible above the horizon at any given time.

More visible satellites means better position accuracy. With exactly four satellites, you can calculate a position fix. With six or more, the receiver can compare multiple redundant solutions and statistically improve accuracy. The “geometry” of the visible satellites also matters—satellites spread widely across the sky give better accuracy than satellites clustered together, because the crossing angles of their distance measurements are more favorable for pinpointing a precise location.

Other countries operate their own global navigation satellite systems (GNSS): Russia’s GLONASS, the European Union’s Galileo, and China’s BeiDou. Modern smartphones typically receive signals from multiple constellations simultaneously, which dramatically increases the number of visible satellites and improves accuracy, particularly in urban environments where some satellites are blocked by buildings.

The Timing Problem—And Why It Requires Relativity

The accuracy of GPS depends entirely on precise timing. A GPS signal travels at the speed of light: 300,000 km/second. An error of 1 microsecond (one millionth of a second) in the timing calculation produces a distance error of 300 meters. For centimeter-level accuracy, timing must be accurate to tens of nanoseconds.

Each GPS satellite carries multiple atomic clocks—cesium or rubidium—accurate to within a nanosecond or so. The clocks on all satellites are synchronized to a common GPS time maintained by the US Air Force, and small corrections are continuously uplinked from ground stations. The satellites also broadcast information about their own clock errors, allowing receivers to correct for any drift.

Here’s where special and general relativity become practically important. GPS satellites are moving at about 14,000 km/h relative to observers on Earth’s surface. According to special relativity, moving clocks run slow—satellite clocks tick about 7 microseconds per day slower than identical clocks on the ground due to their velocity. But the satellites are also at a higher gravitational potential than ground-based clocks; according to general relativity, clocks in weaker gravitational fields run faster—satellite clocks tick about 45 microseconds per day faster than ground clocks due to the reduced gravity at altitude.

The net effect: satellite clocks run about 38 microseconds per day faster than Earth-surface clocks when both effects are combined. Over a day, an uncorrected 38-microsecond error would accumulate to a position error of more than 10 kilometers. GPS satellite clocks are deliberately adjusted to run slightly slower before launch—they’re offset so that once in orbit, the relativistic corrections bring them into sync with ground-based time. Without this relativistic correction, GPS would be useless within minutes. Einstein’s theories aren’t just academic—they’re baked into the operational design of a technology billions of people use daily.

Why Receivers Don’t Need Atomic Clocks

A GPS receiver doesn’t carry an atomic clock—it has only a standard quartz oscillator, which is far less accurate. This creates a problem: the timing measurement that determines satellite distance requires knowing the exact time the signal was received, which requires an accurate receiver clock. How can receivers determine position without atomic clocks?

The fourth satellite measurement solves this. Three satellite distances give you three equations with four unknowns: x, y, z position coordinates, and the receiver’s clock error. A fourth satellite measurement adds a fourth equation, allowing the system to solve for all four unknowns simultaneously—including the clock error. The receiver’s clock doesn’t need to be accurate because its error is treated as an unknown that gets solved as part of the calculation. With four or more satellites, the receiver continuously solves for its own clock error and corrects it. This is why GPS receivers achieve nanosecond-level timing accuracy despite running on cheap quartz oscillators.

Sources of Error and How They’re Corrected

Several factors degrade raw GPS accuracy:

Atmospheric delays: Radio signals slow slightly when passing through the ionosphere (the upper atmosphere, where charged particles interact with radio waves) and the troposphere (the lower atmosphere, where water vapor affects signal propagation). These delays add a few to tens of meters of error. GPS receivers use models of typical atmospheric conditions to estimate and correct for these delays. Dual-frequency receivers (now common in modern smartphones) receive signals at two different frequencies, and since the ionospheric delay varies with frequency, they can directly measure and correct for it.

Multipath errors: In urban environments, satellite signals bounce off buildings before reaching the receiver. The receiver picks up both the direct signal and reflected copies, which can confuse the distance calculation. Modern receivers use signal processing techniques to identify and reject multipath reflections, but urban canyons remain GPS accuracy’s biggest real-world challenge.

Satellite geometry (GDOP): Geometric Dilution of Precision (GDOP) quantifies how satellite positions affect accuracy. Poor satellite geometry amplifies distance measurement errors into larger position errors. Modern receivers select the satellite combination with the best geometry when many satellites are visible.

Ephemeris errors: Each satellite broadcasts its own precise orbital parameters (ephemeris data) so receivers know exactly where the satellite is. Small errors in these parameters translate to position errors. Ground control stations continuously track satellite positions and upload corrections.

Differential GPS and Real-Time Kinematic

For applications requiring accuracy beyond the 2–5 meter level that standard GPS provides, augmentation techniques can push accuracy dramatically further.

Differential GPS (DGPS) uses a network of fixed reference stations at known locations. These stations continuously measure their GPS-calculated position and compare it to their known actual position. The difference reveals the current GPS error in that area (atmospheric conditions and satellite geometry change slowly enough that errors are similar across a region). The error correction is broadcast to nearby receivers, which apply it to their own measurements. DGPS typically achieves sub-meter accuracy.

Real-Time Kinematic (RTK) GPS achieves centimeter-level accuracy by exploiting the carrier wave of the GPS signal (in addition to the coded timing signal). The carrier wave has a wavelength of about 19 cm for L1 GPS. By counting carrier wave cycles and resolving the ambiguity in how many complete cycles occurred between satellite and receiver, RTK can determine distance to within a centimeter. RTK requires a base station within a few tens of kilometers and real-time data communication between base and rover. It’s used in precision agriculture, land surveying, autonomous vehicles, and construction.

Wide-area augmentation systems (WAAS in North America, EGNOS in Europe, MSAS in Japan) broadcast GPS error corrections from geostationary satellites, providing sub-3-meter accuracy across continental areas for aviation and other uses without requiring nearby ground reference stations.

Why Modern Phones Are More Accurate Than Early GPS

Consumer GPS accuracy has improved dramatically over the past decade, driven by several factors:

Dual-frequency reception: GPS Block IIIA satellites and Galileo broadcast a second civilian signal. Smartphones from 2019 onward increasingly receive both L1 and L5 (GPS) or E1 and E5 (Galileo) frequencies, allowing direct correction for ionospheric delays and significantly improving accuracy to the 1–2 meter level.

Multi-constellation support: Modern phones track GPS, GLONASS, Galileo, and BeiDou simultaneously. More satellites means better geometry, faster position fixes, and more reliable coverage in challenging environments.

Better antennas and processing: Chipset improvements and software signal processing have improved multipath rejection and signal tracking.

Assisted GPS (A-GPS): Instead of downloading satellite ephemeris data from the satellites (which takes minutes), phones download it from network servers in milliseconds, enabling fast initial position fixes. A-GPS doesn’t improve steady-state accuracy but eliminates the “cold start” delay that made early GPS receivers frustratingly slow.

What GPS Doesn’t Tell You

GPS tells you where you are horizontally quite well; vertical accuracy (altitude) is typically three to five times worse than horizontal accuracy, because satellite geometry is always better in the horizontal plane—all satellites are above you, none are below. GPS also tells you nothing about what’s around you—the satellite signals don’t carry information about terrain, roads, or buildings. All the mapping intelligence in navigation apps is overlaid from separate map databases on the device or server.

GPS is also susceptible to jamming (overpowering the weak satellite signal with a stronger local signal) and spoofing (transmitting false GPS signals that make receivers calculate incorrect positions). These are real vulnerabilities in high-stakes applications—maritime navigation near certain conflict zones, for example, has been affected by GPS spoofing. Critical applications typically use GPS alongside inertial navigation systems (accelerometers and gyroscopes) that can detect and compensate for GPS anomalies.

None of this diminishes what GPS has achieved: a system designed and built in the 1970s–1980s that remains the foundation of global positioning, navigation, and timing infrastructure for billions of devices, reliably and at no cost to the user. The original constellation has been continuously upgraded, the signals have been modernized, and the technology has been copied and extended by other nations—but the core concept of determining position from satellite timing hasn’t changed since the first GPS satellites launched in 1978.