Technology · 6 min read
How GPS Receivers Calculate Position from Satellite Time Signals
A walkthrough of pseudoranges, four-satellite solutions, atmospheric and relativistic corrections, geometry, and residual error.

How GPS Receivers Calculate Position from Satellite Time Signals
A Global Positioning System receiver does not ask a satellite for its location. It listens to precisely timed radio signals broadcast by multiple satellites, estimates how long each signal traveled, and solves for the receiver's position and clock error. The result is possible because the satellites transmit their orbital data and because radio waves travel at a known speed in vacuum.
That compact description omits most of the engineering. The receiver measures “pseudoranges,” not perfect geometric distances. Satellite clocks, orbital predictions, the atmosphere, reflected signals, and the receiver's own clock all introduce corrections or uncertainty. A position fix is the output of a model fitted to noisy observations.
Signals provide distances with errors attached
Each GPS satellite carries atomic clocks and broadcasts a navigation message along with a coded ranging signal. The message includes the satellite's ephemeris, which describes its predicted orbit, and clock-correction parameters. The receiver generates a replica of the code and shifts it until it aligns with the received pattern. The shift provides an estimate of signal travel time.
Multiplying travel time by the speed of light produces a range-like measurement. A microsecond of timing error corresponds to about 300 meters. Consumer receivers therefore cannot rely on the inexpensive quartz clock in a phone being synchronized to GPS time at atomic-clock precision. Instead, they solve for clock bias along with spatial coordinates.
The term “pseudorange” marks the difference. The observation contains the true satellite-to-receiver distance plus receiver and satellite clock offsets, atmospheric delays, multipath, orbital error, and measurement noise. These terms are modeled with different levels of accuracy. A receiver can still produce a useful position because many errors are predictable, broadcast, estimated, or reduced by additional observations.
Four unknowns require at least four satellites
In three dimensions, the receiver initially has four central unknowns: its x, y, and z coordinates and the offset of its clock from system time. One pseudorange supplies one equation. Observations from at least four suitably placed satellites allow the receiver to solve the four unknowns. Simple explanations often describe three intersecting spheres, but that geometry assumes a synchronized receiver clock; operational GPS normally needs the fourth observation.
The satellite positions and the receiver position are commonly represented in an Earth-centered, Earth-fixed coordinate frame. Because the equations contain distances, they are nonlinear. Receivers start from an approximate location and time, linearize the equations around that estimate, calculate a correction, and iterate. With more than four satellites, a least-squares or filtering method can use redundant observations and estimate uncertainty.
The point is not literally found where four ideal spheres meet. Satellite signals left at different instants while satellites and Earth were moving. The algorithm accounts for transmission time, Earth rotation during propagation, and the chosen reference ellipsoid. Latitude, longitude, and height are derived from the solved coordinates. Elevation relative to mean sea level requires a geoid model; the ellipsoid height reported by the geometric solution is a different quantity.
Corrections turn pseudoranges into a position
Satellite clock behavior is monitored by the GPS control segment, which uploads corrections and fresh orbital parameters. Relativity is built into the system. Special relativity predicts that motion makes orbiting clocks run slower than Earth-fixed clocks, while general relativity predicts that weaker gravity at satellite altitude makes them run faster. The net offset for GPS altitude is about 38 microseconds per day before additional periodic corrections. Satellite frequencies are preset and navigation equations include relativistic terms; otherwise range errors would accumulate rapidly.
The ionosphere delays radio signals by an amount that depends on electron content and frequency. Dual-frequency receivers can estimate much of this delay by comparing signals because the ionospheric effect is dispersive. Single-frequency receivers use a broadcast model, which is less complete. The neutral atmosphere also slows signals; receivers apply models based on elevation angle and atmospheric conditions.
Augmentation and differential techniques improve the remaining estimates. A fixed reference station at a surveyed position can compare observed ranges with expected ones and provide corrections. Satellite-based augmentation systems broadcast integrity information and wide-area corrections. Real-time kinematic positioning uses carrier-phase observations and a nearby reference network to reach centimeter-level results under favorable conditions, but resolving the whole-number carrier-cycle ambiguity is a different and more demanding problem than an ordinary code-based phone fix.
Many current receivers observe more than GPS. Galileo, GLONASS, BeiDou, and regional systems transmit their own signals, orbital data, and system times. Combining constellations increases the number and spread of available satellites, especially where part of the sky is blocked. It also adds parameters: a receiver may estimate an offset between constellation time scales and must use the appropriate reference frames and signal models. “GPS” is often used casually for the displayed location, but the underlying fix may be a multi-constellation global navigation satellite system solution.
Receivers also distinguish code phase from carrier phase. The modulated code identifies a travel-time interval without the severe whole-cycle ambiguity of the much shorter radio carrier. Carrier measurements are more precise but initially reveal only the fractional phase; the unknown integer number of cycles must be resolved using changes over time, multiple satellites, and reference information. This is why survey equipment under clear conditions can outperform a phone even though both hear some of the same satellites.
Accuracy depends on geometry and environment
Having four signals is not enough for a good solution if the satellites occupy nearly the same direction in the sky. Geometric dilution of precision describes how measurement error is amplified by satellite arrangement. Satellites spread across azimuth and elevation generally constrain position better than a tight cluster. Low-elevation signals improve geometry but travel through more atmosphere and are more vulnerable to obstruction, creating a trade-off in receiver selection.
Buildings, cliffs, foliage, and vehicles can block or reflect signals. A reflected path arrives later than the direct path and can bias the pseudorange. Dense urban streets are difficult because the receiver may see few direct signals and many reflections. Modern receivers combine multiple constellations, frequencies, inertial sensors, maps, and statistical filters, but these additions do not make poor observations disappear.
Interference is another boundary. GPS signals reach Earth's surface at low power. Unintentional radio noise and deliberate jamming can prevent tracking; spoofing can present counterfeit signals. Civil receivers can check consistency across frequencies, constellations, motion sensors, or authenticated services where available, but resilience varies by device and threat.
What the receiver knows and what it estimates
GPS positioning rests on established measurements: coded signal timing, broadcast orbits, clock corrections, and a defined terrestrial reference frame. Four or more pseudoranges let a receiver estimate three coordinates and its clock bias. Atmospheric and relativistic corrections are necessary parts of the calculation, not optional refinements to a purely geometric trick.
The uncertainty is local and observable. Accuracy changes with satellite geometry, antenna quality, atmosphere, reflections, interference, and the correction services available. A displayed point is therefore an estimate with an error distribution, even when an interface rounds it to a neat dot. The central premise is supported: satellite time signals can determine position. What remains unknown in any individual fix is the full size and direction of the residual error unless the receiver reports and validates it against independent information.
Sources
4 sources, 1 of them with a link you can follow.
- U.S. Space Force, GPS Standard Positioning Service Performance Standard, 5th edition (2020).
- National Coordination Office for Space-Based Positioning, Navigation, and Timing, GPS.gov pages on trilateration, space segment, and augmentation systems.
- Neil Ashby, “Relativity in the Global Positioning System,” Living Reviews in Relativity 6 (2003), doi:10.12942/lrr-2003-1.
- Bradford Parkinson and James Spilker Jr., eds., Global Positioning System: Theory and Applications, American Institute of Aeronautics and Astronautics (1996).
