Every time you open a map on your phone, something extraordinary happens. Within seconds, a device in your pocket pinpoints your location on the surface of the Earth to within a few meters. It does this by listening to faint whispers from satellites orbiting 20,200 kilometers overhead, doing math that would have been science fiction a century ago, and correcting for the literal warping of time predicted by Einstein’s general theory of relativity.
The Global Positioning System has become so reliable that most people never think about it. But a 2019 study by RTI International for NIST estimated that GPS has generated roughly $1.4 trillion in economic benefits for the U.S. private sector alone since the 1980s, and that losing GPS service would cost approximately $1 billion per day. The system underpins everything from aviation and agriculture to financial trading and emergency response. Understanding how it works is worth the effort.
The Constellation: 31 Satellites, Six Planes, One Purpose
GPS is a U.S.-owned utility that provides positioning, navigation, and timing services to anyone on Earth with a receiver. The system has three segments: space, control, and user. The U.S. Space Force develops, maintains, and operates the space and control segments.
The space segment is a constellation of satellites in medium Earth orbit. According to GPS.gov, these satellites fly at an altitude of approximately 20,200 km (12,550 miles), each circling the Earth twice per day with an orbital period of about 12 hours. They travel at roughly 3.9 km/s, or about 14,000 km/h.
The satellites are arranged into six equally-spaced orbital planes inclined at 55 degrees from the equator. Each plane holds four baseline satellite “slots.” This geometry ensures that at least four satellites are visible from virtually any point on the planet at any time. In practice, the Space Force flies more than the minimum: as of recent counts, 31 operational satellites are in orbit.
Why at least four? Because your GPS receiver needs to solve for four unknowns: your latitude, longitude, altitude, and the exact time. Four satellites give four equations.
How Your Phone Finds You: TrilaterationA method of determining position by measuring distances from at least three known reference points, not angles. GPS uses trilateration, not triangulation., Not Triangulation
GPS does not use triangulation (measuring angles). It uses trilateration (measuring distances). The concept is straightforward, even if the execution is not.
Each satellite continuously broadcasts a signal that contains two key pieces of information: exactly when the signal was sent, and exactly where the satellite was at that moment. Your GPS receiver compares the time signals it receives from the currently visible satellites and calculates the distance to each one. Since radio signals travel at the speed of light (about 299,792,458 meters per second), multiplying the travel time by the speed of light gives the distance.
With a distance from one satellite, you know you are somewhere on a sphere centered on that satellite. Two satellites narrow it to the intersection of two spheres (a circle). Three satellites narrow it to two points. One of those points is usually absurd (deep in space or inside the Earth), so three satellites can theoretically give you a position.
But there is a catch. Your phone does not carry an atomic clock. Its internal clock has significant error. So the system uses a fourth satellite to solve for the clock error simultaneously with your position. That is why four is the magic number.
Atomic Clocks: Why a Billionth of a Second Matters
The entire system rests on timing. Light travels about 30 centimeters in one nanosecond (one billionth of a second). If the satellite clocks are off by even a few nanoseconds, your position can be off by meters. As physicist Neil Ashby wrote in Physics Today, if navigation errors of more than a meter are to be avoided, an atomic clock must deviate by less than about 4 nanoseconds from perfect synchronization with the other satellite clocks. Only atomic clocks can do that.
Each GPS satellite carries multiple atomic clocks. The fundamental timekeeper is the cesium-133 atom, whose outer electron oscillates between two energy states at a very specific frequency. According to NIST, you count 9,192,631,770 cycles of microwave radiation tuned to cesium’s natural resonant frequency, and the time it takes to count those cycles equals exactly one second. This is not arbitrary: the number was established by comparing cesium resonance to astronomical observations of the year’s length in the late 1950s, and in 1967, the world officially redefined the second based on it.
Modern GPS satellites carry rubidium atomic clocks that are smaller and lighter than cesium beam clocks, though ground stations maintain cesium standards. The U.S. Naval Observatory’s ensemble of about 50 cesium-beam frequency standards and a dozen hydrogen masers provides the reference for GPS time.
Einstein to the Rescue: Why Relativity Is Not Optional
Here is where GPS gets genuinely remarkable. The system would be useless without corrections from both of Einstein’s theories of relativity.
Special relativity (time dilation): The satellites move at about 3.9 km/s relative to observers on the ground. According to special relativity, moving clocks tick more slowly. This causes the satellite clocks to fall behind ground clocks by about 7 microseconds per day.
General relativity (gravitational time dilation): The satellites orbit at 20,200 km altitude, far from Earth’s gravitational well. General relativity predicts that clocks farther from a massive object tick faster. The satellite clocks gain about 45 microseconds per day relative to ground clocks.
These two effects work in opposite directions, but they do not cancel. The net result is that satellite clocks tick faster than ground clocks by about 38 microseconds per day (45 minus 7). That is 38,000 nanoseconds. Left uncorrected, this would cause positioning errors accumulating at a rate of about 10 to 11 kilometers per day.
The engineers who designed GPS accounted for this. Before launch, each satellite’s atomic clock is deliberately set to tick at a slightly lower frequency: 10.22999999543 MHz instead of the nominal 10.23 MHz. This pre-correction compensates for the general relativistic speedup once the satellite reaches orbit. The receiver then handles additional special relativistic corrections in real time, using orbital data transmitted by the satellites.
As Ashby wrote in Living Reviews in Relativity, these clocks “have gravitational and motional frequency shifts which are so large that, without carefully accounting for numerous relativistic effects, the system would not work.”
What Can Go Wrong: Error Sources
Even with perfect clocks and relativistic corrections, several factors degrade GPS accuracy:
- The ionosphere. The ionosphere is the biggest source of GPS error. GPS signals pass through this charged layer of the upper atmosphere, where free electrons scatter and delay them by nanoseconds to microseconds. Even a few nanoseconds of delay can cause meters of positioning error. The ionosphere’s behavior varies with solar activity, time of day, and season, making it difficult to predict.
- MultipathA GPS error caused when signals bounce off buildings, terrain, or water before reaching the receiver, arriving later than the direct-path signal. Common in cities.. Multipath occurs when a signal bounces off buildings, trees, water, or other surfaces before reaching the antenna. The reflected signal arrives later than the direct signal, introducing timing errors. This is why GPS accuracy drops in dense urban areas and forests.
- Satellite geometry. If the visible satellites are clustered in one part of the sky rather than spread out, the math becomes less precise. This “dilution of precisionA measure of how satellite geometry affects GPS accuracy. When visible satellites are clustered together rather than spread across the sky, position errors are amplified.” (DOP) worsens in canyons, valleys, or anywhere that limits the receiver’s view of the sky.
- Tropospheric delay. The lower atmosphere also slows GPS signals slightly, adding 2.5 to 25 meters of range error depending on satellite elevation angle.
From Military Secret to Everyday Utility
GPS was developed by the U.S. Department of Defense as a military navigation system. For years, the military deliberately degraded the civilian signal through a policy called Selective Availability (SA), which introduced intentional clock errors that limited civilian accuracy to about 100 meters.
In May 2000, at the direction of President Bill Clinton, the U.S. government turned off Selective Availability permanently. Overnight, civilian GPS accuracy improved roughly tenfold. The United States has stated it has no intent to ever use Selective Availability again, and GPS III satellites are being built without the SA capability entirely.
Today, a basic civilian GPS receiver can determine your position to within about 5 to 10 meters. More advanced techniques like Differential GPS (DGPS) and Real-Time Kinematic (RTK) methods deliver centimeter-level positions, enabling precision agriculture, autonomous vehicles, and high-accuracy surveying.
The Next Generation: New Signals, Better Accuracy
GPS is not standing still. The modernization program is adding three new civilian signals: L2C (for ionospheric correction and faster acquisition), L5 (for safety-of-life transportation applications with higher power and bandwidth), and L1C (for interoperabilityThe ability of military forces or equipment from different nations to function together seamlessly in joint operations. with international systems like Galileo and BeiDou).
Through a technique called trilaning, using three GPS frequencies simultaneously may enable sub-meter accuracy without any augmentation systems. L5, broadcasting in a radio band reserved exclusively for aviation safety, will provide the most advanced civilian GPS signal yet.
These signals are phasing in incrementally as newer satellites replace older ones. The full constellation of modernized satellites is expected to be in place by the late 2020s.
Why This Matters
GPS is one of the clearest examples of fundamental physics having direct, practical consequences for billions of people. The system requires atomic clocks that count 9.2 billion oscillations per second with nanosecond accuracy. It requires orbital mechanics that place satellites in precise geometries around the planet. And it requires corrections from both special and general relativity, theories that most people associate with black holes and thought experiments, not with finding the nearest grocery store.
The next time your phone tells you to turn left in 200 meters, consider what just happened. At least four satellites, each carrying atomic clocks tuned to compensate for the warping of spacetime, transmitted signals at the speed of light. Those signals crossed 20,200 kilometers of space and atmosphere, were scattered by the ionosphere and reflected off buildings, and arrived at your phone within a few nanoseconds of the predicted time. Your phone solved four simultaneous equations, applied relativistic corrections, and plotted your position on a map. The whole process took seconds.
It is, by any measure, one of the most elegant engineering achievements in human history. It works so well that no one notices it is there.
The Global Positioning System is a masterwork of applied physics that fuses atomic timekeeping, orbital mechanics, electromagnetic signal processing, and relativistic corrections into a system that delivers meter-level positioning to billions of receivers worldwide. A 2019 NIST-commissioned study valued GPS at $1.4 trillion in U.S. private-sector economic benefits since the 1980s, with an estimated $1 billion per day cost if the system were lost. Understanding the physics beneath the interface is worth the effort.
The Space Segment: Orbital Architecture
The GPS constellation consists of at least 24 satellites (31 operational as of recent counts) in medium Earth orbit at an altitude of approximately 20,200 km (12,550 miles). The satellites are distributed across six orbital planes, each inclined at 55 degrees to the equatorial plane, with four baseline slots per plane. This geometry guarantees that at least four satellites are above the horizon from any point on Earth at any given time, with typical visibility of 6 to 12 satellites.
Each satellite completes one orbit in approximately 11 hours and 58 minutes (half a sidereal day), at an orbital velocity of roughly 3.87 km/s. This is not a geostationary orbitAn orbit at ~35,786 km altitude where a satellite's orbital period matches Earth's rotation, making it appear stationary above one point.; a fixed observer sees the same satellite at nearly the same position on the celestial sphere twice per sidereal day.
PseudorangeA GPS receiver's estimated distance to a satellite, calculated from signal travel time but not yet corrected for the receiver's clock error. Measurement and TrilaterationA method of determining position by measuring distances from at least three known reference points, not angles. GPS uses trilateration, not triangulation.
The fundamental principle of GPS navigation is an application of the constancy of the speed of light. Each satellite transmits a coded timing signal. The receiver measures the time delay between transmission and reception, multiplies by c (299,792,458 m/s, defined exactly), and obtains a pseudorange to each satellite.
The term “pseudorange” (not “range”) is critical. The receiver’s internal quartz oscillator has substantial clock bias compared to the atomic clocks on the satellites. This bias is an additional unknown. With four unknowns (three spatial coordinates plus clock bias), the receiver needs signals from at least four satellites. The system of equations takes the form:
|r - r_i| = c(t - t_i), for i = 1, 2, 3, 4
where r and t are the receiver’s position and clock time, and r_i and t_i are the satellite positions and transmission times. Solving this nonlinear system of equations (typically via iterative least-squares or Kalman filtering, often using more than four satellites for overdetermined solutions) yields the position fix.
Atomic Timekeeping: The Cesium Standard
GPS accuracy is fundamentally limited by timing precision. Light travels approximately 30 cm per nanosecond, so meter-level navigation requires clock synchronization to within about 4 nanoseconds, a fractional time stability better than 1 part in 1013. Only atomic clocks achieve this.
The SI second is defined by the cesium-133 atom: exactly 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state. This number was established in the late 1950s by comparing cesium resonance measurements at the UK’s National Physical Laboratory with astronomical observations at the U.S. Naval Observatory, and became the official definition in 1967.
Current GPS satellites carry rubidium atomic frequency standards (smaller and lighter than cesium beam clocks), with typical stability of a few parts in 1014 over a day. The ground segment reference is the U.S. Naval Observatory’s clock ensemble (approximately 50 cesium-beam standards and a dozen hydrogen masers), which maintains GPS time to within about 20 ns of UTC(USNO).
Relativistic Corrections: The Physics That Makes It Work
As Neil Ashby documented in Living Reviews in Relativity, GPS clocks have gravitational and motional frequency shifts so large that, without carefully accounting for numerous relativistic effects, the system would not work. The corrections involve both special and general relativity.
Special Relativistic Time Dilation
At orbital velocity v of approximately 3.87 km/s, the second-order Doppler shift causes the satellite clocks to run slow by a fraction v2/2c2. This amounts to about 7 microseconds per day lost relative to ground clocks.
General Relativistic Gravitational Blueshift
At 20,200 km altitude, the gravitational potential is weaker than at Earth’s surface. General relativity predicts that clocks in weaker gravitational fields tick faster, causing the satellite clocks to gain approximately 45 microseconds per day relative to ground clocks.
Net Effect and Compensation
The net relativistic drift is approximately +38 microseconds per day (45 – 7 = 38), or 38,000 nanoseconds. Uncorrected, this would accumulate positioning errors of about 11.4 km per day.
The primary correction is applied before launch. The satellite oscillators are set to a frequency of 10.22999999543 MHz rather than the nominal 10.23 MHz, a fractional offset of approximately 4.47 x 10-10. This compensates for the net general relativistic frequency shift once the satellite reaches its operational orbit. Residual special relativistic corrections (which depend on the satellite’s specific orbital parameters and eccentricity) are computed by the receiver in real time using ephemeris data broadcast by the satellites.
Additional relativistic effects accounted for in the system include the Sagnac effect (due to Earth’s rotation, contributing up to 207 nanoseconds for equatorial paths) and eccentricity corrections (satellite orbits are not perfectly circular, causing periodic relativistic oscillations of up to 46 nanoseconds).
Error Budget: What Degrades the Fix
Several physical phenomena introduce errors beyond clock and relativistic corrections:
- Ionospheric delay. The ionosphere is the single largest source of GPS error. Free electrons in the upper atmosphere (60 to 1,000 km altitude) cause frequency-dependent signal delays. Single-frequency receivers use broadcast ionospheric models (the Klobuchar model corrects roughly 50% of the delay). Dual-frequency receivers exploit the dispersive nature of the ionosphere: since the delay is inversely proportional to frequency squared, measuring the same signal at L1 (1575.42 MHz) and L2 (1227.60 MHz) allows direct computation and removal of the ionospheric term.
- Tropospheric delay. The neutral atmosphere (0 to ~12 km altitude) introduces a non-dispersive delay of approximately 2.5 to 25 meters of additional range, depending on satellite elevation angle. Standard models (Saastamoinen, Hopfield) correct for most of this, but residual errors of 0.2 to 1 meter can persist, particularly in the wet component (water vapor).
- MultipathA GPS error caused when signals bounce off buildings, terrain, or water before reaching the receiver, arriving later than the direct-path signal. Common in cities.. Reflected signals arriving via indirect paths from buildings, terrain, or water surfaces corrupt the pseudorange measurement. Multipath errors can range from centimeters to meters and are a primary concern in urban canyons and near reflective surfaces.
- Dilution of PrecisionA measure of how satellite geometry affects GPS accuracy. When visible satellites are clustered together rather than spread across the sky, position errors are amplified. (DOP). The geometric arrangement of visible satellites directly affects position uncertainty. When satellites are clustered rather than well-distributed across the sky, the same pseudorange errors translate into larger position errors. The 55-degree orbital inclination of GPS was chosen partly to optimize satellite geometry for mid-latitude users.
- Satellite ephemeris and clock errors. Despite continuous monitoring by the Control Segment, residual errors in the broadcast orbit and clock parameters contribute approximately 1 to 2 meters of ranging error.
Signal Structure and Modernization
The legacy civilian signal (L1 C/A) broadcasts a 1.023 MHz pseudorandom noise code on the L1 carrier at 1575.42 MHz. This frequency is an integral multiple of the fundamental 10.23 MHz clock frequency. The C/A code repeats every millisecond, giving a “chip” length of about 300 meters, which sets the theoretical pseudorange resolution (though carrier-phase measurements achieve millimeter-level precision).
The GPS modernization program is adding three new civilian signals:
- L2C (1227.60 MHz): Dedicated civilian signal on L2, enabling ionospheric correction for dual-frequency receivers. Broadcasting from 25 satellites as of mid-2023. Provides higher effective power than L1 C/A.
- L5 (1176.45 MHz): Safety-of-life signal in a protected aeronautical band, with greater bandwidth and power for enhanced jam resistance. Broadcasting from 18 satellites as of mid-2023, with full constellation expected around 2027.
- L1C (1575.42 MHz): MBOC-modulated signal designed for interoperabilityThe ability of military forces or equipment from different nations to function together seamlessly in joint operations. with Galileo, BeiDou, and QZSS. Broadcasting from 6 satellites as of mid-2023.
Through trilaning (combining three frequencies), sub-meter accuracy may be achievable without any external augmentation. This represents a fundamental shift: the same physics that required $50,000 military receivers in the 1990s will deliver centimeter-class positioning to consumer devices.
From Selective Availability to Open Architecture
Until May 2000, the DoD imposed Selective Availability: intentional dithering of satellite clock and ephemeris data that limited civilian accuracy to approximately 100 meters (95% horizontal). President Clinton ordered SA turned off on May 1, 2000, immediately improving civilian accuracy by roughly an order of magnitude. The U.S. government has stated that regional denial capabilities provide adequate military advantage without global signal degradation. GPS III satellites are being built without SA capability, making the decision permanent.
Current civilian accuracy with a basic single-frequency receiver is approximately 5 to 10 meters. DGPS narrows this to 1 to 3 meters. RTK, using carrier-phase measurements from a nearby reference station, achieves 1 to 2 centimeters. PPP (Precise Point Positioning), using global correction data, can reach similar accuracy without local infrastructure.
The Deeper Point
GPS is arguably the most sophisticated piece of physics that billions of people use daily without understanding. The system simultaneously depends on Newtonian gravity (orbital mechanics), quantum mechanics (atomic clocks exploiting hyperfine transitions in cesium-133), special relativity (velocity-dependent time dilation), general relativity (gravitational frequency shifts), and electromagnetic wave propagation through a dispersive, turbulent atmosphere.
Remove any one of these, and the system fails. Skip the relativistic corrections, and positioning errors exceed 11 km per day. Use a quartz oscillator instead of an atomic clock, and timing drift renders the fix meaningless within minutes. Ignore the ionosphere, and your position wanders by meters unpredictably.
The next time your phone resolves your position in seconds, recognize that what just happened is not simple. At least four satellites, each carrying atomic clocks pre-corrected for the curvature of spacetime, transmitted signals that crossed 20,200 km of vacuum and atmosphere. Your receiver decoded pseudorandom noise codes, solved a nonlinear system of equations, applied relativistic and atmospheric corrections, and plotted the result on a map. The entire process, from signal transmission to rendered position, took less time than it takes to read this sentence.
That is not convenience. That is physics, working exactly as predicted.



