On February 22, 1978, the first Navstar GPS satellite lifted off from Vandenberg Air Force Base. The engineers who built it had solved a problem that seemed impossible: determining a position anywhere on Earth to within meters, using signals from satellites orbiting 20,000 kilometers away. The solution required not just advances in electronics and rocketry, but a practical application of Einstein’s theory of relativity that affects every GPS receiver in existence today.
Most people know GPS involves satellites and can pinpoint location. Fewer understand that the system fundamentally measures time, not distance. A GPS receiver doesn’t “see” satellites or measure angles. It listens for precisely timed radio signals, calculates how long they took to arrive, and works backward to determine position through pure mathematics. This approach—trilateration—requires timing accuracy measured in nanoseconds. A clock error of just 10 nanoseconds translates to a position error of about 3 meters.
The system that emerged from decades of military development now guides billions of smartphones, enables precision agriculture, synchronizes financial markets, and times cellular networks. Understanding how GPS works reveals engineering at the intersection of orbital mechanics, quantum physics, and relativity.
The Geometry of Position: Trilateration
GPS does not use triangulation. This distinction matters because triangulation measures angles, while GPS measures distances. Trilateration—determining position from measured distances to known points—is the actual technique.
Consider how this works in two dimensions. If you know you’re exactly 20 kilometers from Tower A, your position could be anywhere on a circle with a 20-kilometer radius centered on Tower A. Add a second distance—say, 30 kilometers from Tower B—and your possible positions narrow to the two points where these circles intersect. A third measurement eliminates the ambiguity, pinpointing your exact location.

Image source: GIS Geography - How GPS Receivers Work: Trilateration vs Triangulation
GPS extends this principle to three dimensions. Each satellite broadcasts its position and the time its signal was transmitted. The receiver measures the time delay and calculates distance using the speed of light:
$$d = c \times \Delta t$$where $c$ is the speed of light (approximately 299,792,458 meters per second) and $\Delta t$ is the travel time.
With one satellite, your position could be anywhere on a sphere with radius equal to the measured distance. A second satellite adds another sphere, and their intersection forms a circle. A third sphere intersects this circle at two points—one on or near Earth’s surface, the other in space. A fourth satellite resolves this ambiguity.
The fourth satellite serves another critical purpose: it corrects for clock error. GPS receivers contain inexpensive quartz oscillators, not atomic clocks. The receiver’s clock may differ from GPS time by milliseconds, introducing a common bias to all distance measurements. By solving for four unknowns—three position coordinates plus clock bias—a receiver can determine position without needing an atomic clock of its own.
The Mathematics Behind the Fix
The GPS position solution involves solving a system of nonlinear equations. For each satellite $i$, the pseudorange measurement $\rho_i$ relates to the receiver’s position $(x, y, z)$ and clock bias $b$:
$$\rho_i = \sqrt{(x - x_i)^2 + (y - y_i)^2 + (z - z_i)^2} + c \cdot b$$where $(x_i, y_i, z_i)$ is the satellite’s known position. With four or more satellites, the receiver solves this system iteratively, typically using least-squares estimation. The term “pseudorange” acknowledges that these measurements include the receiver clock bias—hence “pseudo” rather than true range.
The Constellation: Twenty-Four Clocks in the Sky
The GPS constellation nominally consists of 24 satellites distributed across six orbital planes, each inclined at 55 degrees to the equator. The satellites orbit at approximately 20,200 kilometers altitude with a period of 11 hours and 58 minutes—half a sidereal day. This period ensures each satellite passes over the same ground track twice daily.
mindmap
root((GPS Constellation))
Orbital Configuration
6 orbital planes
55 degree inclination
20,200 km altitude
11h 58m orbital period
Satellite Distribution
4 satellites per plane
60 degree plane separation
24+ operational satellites
31 typically active
Coverage
4-12 satellites visible
Global coverage
Any point on Earth
24/7 availability
The constellation design ensures that between 4 and 12 satellites are visible from any point on Earth at any time. More visible satellites generally mean better position accuracy because the geometry improves—a concept quantified as Dilution of Precision (DOP).
Satellites continuously broadcast their position (calculated from orbital parameters called ephemeris data) and the precise time. Ground monitoring stations track each satellite, update its orbital parameters, and correct any clock drift. This information uploads to satellites several times daily.
Atomic Clocks: The Heart of GPS
Every GPS satellite carries multiple atomic clocks—typically a combination of cesium and rubidium standards. These clocks measure time by counting the oscillations of atoms transitioning between energy states. A cesium clock counts 9,192,631,770 oscillations per second by definition; a rubidium clock operates at approximately 6,834,682,611 Hz.
The stability requirement is extraordinary. A GPS satellite clock must maintain accuracy within about 10 nanoseconds over a day—a stability of roughly one part in $10^{13}$. This means if you ran such a clock for 300,000 years, it would drift by only about one second.
The United States Naval Observatory maintains the master clock for GPS, using an ensemble of hydrogen masers, cesium clocks, and rubidium clocks. Ground stations compare each satellite’s clock against this master and upload correction parameters that receivers use to adjust the received time signals.
Why Such Precision Matters
Light travels approximately 30 centimeters per nanosecond. A clock error of one microsecond (1,000 nanoseconds) produces a range error of 300 meters. GPS requires nanosecond-level timing because meter-level positioning demands it. The system specification calls for position accuracy of about 3-5 meters for civilian users, which translates to timing accuracy of about 10-15 nanoseconds.
The Relativity Correction: Einstein in Your Pocket
GPS would fail without accounting for Einstein’s theories of relativity. This isn’t a subtle effect—it’s a dominant one, amounting to about 38 microseconds per day of clock drift.
Special Relativity: Moving Clocks Run Slow
Special relativity predicts that moving clocks tick more slowly than stationary ones. GPS satellites orbit at about 3.9 kilometers per second relative to Earth’s center. At this velocity, time dilation causes satellite clocks to lose about 7 microseconds per day compared to clocks on the ground.
General Relativity: Gravity Slows Time
General relativity predicts that clocks in stronger gravitational fields tick more slowly. GPS satellites orbit at 20,200 kilometers altitude where Earth’s gravitational field is weaker than at the surface. This causes satellite clocks to run faster than ground clocks by about 45 microseconds per day.
The Net Effect and Its Correction
The two effects combine: -7 microseconds (special relativity) + 45 microseconds (general relativity) = +38 microseconds per day. Without correction, satellite clocks would gain 38 microseconds daily, accumulating a timing error of about 11 kilometers per day in range measurements.
The solution is elegant. Before launch, each satellite clock is set to run slightly slow—specifically, at 10.22999999543 MHz instead of the nominal 10.23 MHz. This “factory offset” compensates for the average relativistic effect, so the clock appears to tick at the correct rate once in orbit.
However, satellite orbits aren’t perfectly circular. Orbital eccentricity causes the satellite’s speed and altitude to vary slightly throughout each orbit, producing a periodic relativistic correction of up to about 45 nanoseconds (roughly 15 meters of range error). Receivers calculate this correction based on the satellite’s precise orbital position.
graph LR
A[Special Relativity<br/>-7 μs/day] --> C[Net Effect<br/>+38 μs/day]
B[General Relativity<br/>+45 μs/day] --> C
C --> D[Factory Offset<br/>10.22999999543 MHz]
D --> E[Corrected Clock<br/>in Orbit]
style A fill:#ffcccc
style B fill:#ccffcc
style C fill:#ffffcc
style E fill:#ccffcc
The Signal: Finding Needles in a Haystack
GPS satellites transmit on multiple frequencies, but the primary civilian signal broadcasts at 1575.42 MHz (designated L1). Each satellite transmits a unique code that allows receivers to distinguish signals from different satellites despite them sharing the same frequency—a technique called Code Division Multiple Access (CDMA).
Gold Codes and Pseudo-Random Noise
The civilian Coarse/Acquisition (C/A) code is a sequence of 1,023 bits called a Gold code, named after Robert Gold who discovered this class of sequences in 1967. Each satellite is assigned a unique 1023-bit code that repeats every millisecond. The codes are designed to have very low cross-correlation—the mathematical measure of how much two sequences resemble each other—allowing receivers to separate signals from different satellites.
The signal arrives at the receiver with a power of about -130 dBm—roughly 0.1 femtowatts, far below thermal noise. Yet receivers extract this signal through correlation. The receiver generates a local copy of the expected code and shifts it in time until correlation peaks, indicating alignment with the incoming signal. This time shift, combined with the known transmission time embedded in the signal, yields the travel time measurement.
Signal Structure
The L1 signal carries:
- C/A code: The civilian ranging code, transmitted at 1.023 million bits per second
- Navigation message: Satellite position, clock corrections, and system status at 50 bits per second
- P(Y) code: A precision encrypted military code transmitted at 10.23 million bits per second
Modern GPS satellites also broadcast on L2 (1227.60 MHz) and L5 (1176.45 MHz) frequencies, enabling dual-frequency receivers to correct for ionospheric delay by comparing how the two frequencies propagate through the ionized upper atmosphere.
Error Sources: Why GPS Isn’t Perfect
Even with perfect clocks and relativity corrections, GPS accuracy degrades from multiple error sources:
| Error Source | Typical Magnitude |
|---|---|
| Ionospheric delay | ±5 meters |
| Satellite clock error | ±2 meters |
| Orbit (ephemeris) error | ±2.5 meters |
| Multipath | ±1 meter |
| Tropospheric delay | ±0.5 meters |
| Receiver noise | ±0.3 meters |
Atmospheric Effects
The ionosphere—a layer of charged particles from 50 to 1,000 kilometers altitude—slows radio signals in a frequency-dependent manner. Single-frequency receivers estimate this delay using mathematical models, but models can’t capture real-time variations caused by solar activity. Dual-frequency receivers measure ionospheric delay directly by comparing propagation times at two frequencies, significantly improving accuracy.
The troposphere—the lower atmosphere where weather occurs—also delays signals, primarily due to water vapor content. Tropospheric delay is more difficult to model but generally smaller than ionospheric effects.
Multipath
When GPS signals reflect off buildings, terrain, or other surfaces before reaching the receiver, the longer path length creates measurement errors. Modern receivers employ sophisticated algorithms to detect and reject multipath signals, but in urban canyons or near large reflective surfaces, multipath remains a significant accuracy limitation.
Satellite Geometry: Dilution of Precision
The arrangement of visible satellites affects accuracy. If all visible satellites cluster in one part of the sky, position accuracy degrades. The geometric measure of this effect is Dilution of Precision (DOP), expressed as:
- HDOP: Horizontal DOP—affects latitude/longitude accuracy
- VDOP: Vertical DOP—affects altitude accuracy (typically worse than horizontal)
- PDOP: Position DOP—combines horizontal and vertical
- TDOP: Time DOP—affects clock accuracy
- GDOP: Geometric DOP—combines all factors
A PDOP of 2 is considered good; values above 6 indicate poor geometry. Receivers often report these values, and some applications set DOP thresholds below which position fixes are rejected.
Augmentation: Pushing Accuracy Further
Standard GPS provides meter-level accuracy. Several augmentation systems improve this to centimeter-level for applications requiring precision:
SBAS: Satellite-Based Augmentation Systems
The Wide Area Augmentation System (WAAS) and similar systems elsewhere broadcast correction data from geostationary satellites. WAAS provides corrections for satellite clock errors, orbit errors, and ionospheric delays across North America. WAAS-enabled receivers typically achieve 1-3 meter accuracy.
RTK: Real-Time Kinematic
RTK uses carrier phase measurements—the phase of the GPS carrier wave—to achieve centimeter accuracy. A base station at a known location transmits correction data to a rover receiver in real-time. RTK requires relatively short baselines (typically under 20 kilometers) and dual-frequency receivers.
PPP: Precise Point Positioning
PPP uses precise satellite orbit and clock data from global networks to achieve decimeter or better accuracy without a local base station. The trade-off is convergence time—it can take 20-30 minutes for a PPP solution to reach full accuracy.
The Invisible Infrastructure
GPS succeeds through layers of invisible infrastructure. Twenty-four satellites broadcast signals that any receiver can use. Ground stations continuously monitor satellites and upload corrections. The system maintains its own time scale—GPS Time—continuous since January 6, 1980, unburdened by the leap seconds that complicate UTC.
The economics are remarkable. The United States spends roughly $2 million daily operating GPS, yet the system provides free positioning and timing to an estimated 6 billion receivers worldwide. Every smartphone, most vehicles, countless industrial systems, and critical infrastructure depend on signals that arrive with less power than a firefly’s glow.
This dependence creates vulnerabilities. GPS signals are weak enough to be jammed by inexpensive transmitters. Spoofing—transmitting fake GPS signals—can deceive receivers into reporting false positions. In 2011, Iran claimed to have captured a drone by spoofing its GPS. Multiple studies have demonstrated GPS spoofing against ships, aircraft, and financial markets.
Beyond GPS: The GNSS Constellation
GPS isn’t alone. Russia operates GLONASS, the European Union provides Galileo, China has BeiDou, and Japan and India operate regional systems. Modern receivers typically use multiple constellations simultaneously—multiconstellation GNSS improves availability, accuracy, and resilience.
Each system makes different engineering choices. GLONASS uses Frequency Division Multiple Access (FDMA) rather than CDMA, requiring receivers to handle multiple frequencies. Galileo broadcasts more precise signals and offers a commercial encrypted service. BeiDou includes satellites in geostationary orbit for regional augmentation.
What Your Phone Actually Does
When you open a mapping application, your phone executes a sequence that would have seemed like science fiction in 1978:
- Acquisition: The GPS receiver searches for signals from any satellite in view, correlating against known PRN codes
- Tracking: Once acquired, the receiver continuously tracks each satellite’s signal, maintaining code and carrier phase lock
- Decoding: The receiver extracts navigation data from each signal—satellite position, clock corrections, system status
- Position calculation: Using pseudoranges from four or more satellites, the receiver solves the trilateration equations
- Output: Position, velocity, and time estimates update several times per second
This entire process typically completes within seconds of a cold start, with ongoing position updates at 1-10 Hz. The receiver also corrects for relativistic effects, ionospheric delay (using models or dual-frequency measurements if available), and satellite clock errors. All from signals that began their journey 20,000 kilometers away, 67 milliseconds before you read them.
The Achievement
GPS represents one of the most successful engineering projects in history. What began as a military navigation system has become invisible infrastructure—a utility so reliable and ubiquitous that most users never consider what makes it work. The system applies special and general relativity in a practical device, maintains nanosecond timing across a global network, and delivers meter-level positioning to anyone with a receiver.
The engineers who designed GPS solved problems that spanned disciplines: orbital mechanics for satellite positioning, quantum physics for atomic clocks, signal processing for code correlation, and relativistic physics for time corrections. The result is a system that works so well it disappears into the background of modern life—exactly what great infrastructure should do.
References
-
Parkinson, B. W. (1996). GPS Error Analysis. In Global Positioning System: Theory and Practice (pp. 469-483). American Institute of Aeronautics and Astronautics.
-
Ashby, N. (2003). Relativity in the Global Positioning System. Living Reviews in Relativity, 6(1), 1. https://link.springer.com/article/10.12942/lrr-2003-1
-
GPS World. (2023). Inside the box: GPS and relativity. https://www.gpsworld.com/inside-the-box-gps-and-relativity/
-
Kaplan, E. D., & Hegarty, C. J. (2005). Understanding GPS: Principles and Applications (2nd ed.). Artech House.
-
NovAtel. (2024). Chapter 4: GNSS Error Sources. https://novatel.com/an-introduction-to-gnss/gnss-error-sources
-
Let’s Talk Science. (2024). The Math Behind GPS. https://letstalkscience.ca/educational-resources/backgrounders/math-behind-gps
-
GIS Geography. (2024). How GPS Receivers Work - Trilateration vs Triangulation. https://gisgeography.com/trilateration-triangulation-gps/
-
All About Circuits. (2017). GPS Times, Atomic Clock Frequencies, and the Increasing Accuracy of GPS. https://www.allaboutcircuits.com/technical-articles/gps-times-atomic-clock-frequencies-increasing-accuracy-of-gps/
-
Natronics. (2014). The GPS PRN (Gold Codes). https://natronics.github.io/blag/2014/gps-prn/
-
NOAA. (2024). The Global Positioning System. https://oceanservice.noaa.gov/education/tutorial_geodesy/geo09_gps.html
-
NASA. (2023). GPS. https://www.nasa.gov/directorates/somd/space-communications-navigation-program/gps/
-
Wikipedia. Global Positioning System. https://en.wikipedia.org/wiki/Global_Positioning_System
-
Wikipedia. GPS Signals. https://en.wikipedia.org/wiki/GPS_signals
-
Navipedia. GPS Space Segment. https://gssc.esa.int/navipedia/index.php/GPS_Space_Segment
-
Aerospace Corporation. Brief History of GPS. https://aerospace.org/article/brief-history-gps
-
GPS.gov. Pseudorandom Noise Code Assignments. https://www.gps.gov/pseudorandom-noise-code-assignments
-
SBG Systems. GNSS and their Error Sources. https://www.sbg-systems.com/technology/gnss-and-their-error-sources/
-
GNSS Decoded. Dilution of Precision - PDOP, HDOP, VDOP, TDOP, GDOP. https://gnssdecoded.com/dilution-of-precision/