How GPS Calculates Your Position: Geometry and Relativity

Why does a satellite orbiting 20,000 kilometers overhead need to know about Einstein just to tell your phone which street you're on?

You'd think GPS was pure geometry. Measure some distances, do some algebra, done. And for the first half of the story, that's exactly right. But the second half is stranger. It turns out the clocks on those satellites tick at a genuinely different rate than the clock in your pocket, for reasons that have nothing to do with engineering and everything to do with the physics of space and time itself.

Let's build the whole thing, geometry first, then the relativity nobody expects.

The geometry: how many circles do you need?

Picture a satellite broadcasting a signal. Your phone catches it and measures how long the signal took to arrive. Multiply that travel time by the speed of light and you get a distance, call it $r$. That's one number. And here's the thing: one distance from one satellite doesn't tell you where you are. It tells you that you're somewhere on a circle of radius $r$ around that satellite.

Add a second satellite. Now you have a second circle. Two circles cross at exactly two points, so your options just dropped from infinite to two.

Add a third satellite, a third circle, and something clean happens: three circles in general position intersect at exactly one point. That point is you.

Drag the slider above and watch it happen live. One satellite gives you a whole ring of possibilities. Two narrows it to a pair of candidate points. Three collapses it to a single location, no ambiguity left. This trick is called trilateration, and it's the geometric engine underneath every GPS fix.

Algebraically, each circle is an equation, $(x - x_i)^2 + (y - y_i)^2 = r_i^2$. Subtract any two of these equations and the squared terms cancel, leaving a linear equation in $x$ and $y$. Do that twice and you've got two linear equations, two unknowns, solvable in a few lines of algebra. The hard part was never the algebra.

The hard part is measuring $r$ correctly

Distance comes from time: $r = c \cdot \Delta t$, where $c$ is the speed of light and $\Delta t$ is how long the signal took to arrive. Light is fast, about 300,000 km per second, which means a timing error of just one microsecond turns into a 300 meter position error. GPS lives and dies by how precisely it can measure time.

So satellites carry atomic clocks. And here's where the story stops being an engineering problem and turns into a physics one, because it turns out that a clock's rate depends on more than how well it's built. It depends on how fast the clock is moving, and how deep it sits in a gravitational field. Both of those apply to a GPS satellite, and both were predicted by Einstein decades before the first satellite ever launched.

Special relativity: a moving clock runs slow

Here's the first effect, and it's not a metaphor. Any clock in motion, relative to you, ticks slower than an identical clock at rest next to you. It's a real, measured, repeatedly confirmed consequence of special relativity. The faster the motion, the bigger the slowdown.

GPS satellites orbit at roughly 14,000 km/h, which is fast enough for this effect to matter. But here's a wrinkle worth noticing: a higher orbit doesn't mean a faster satellite. It's the opposite. Gravity gets weaker with distance, so a satellite farther out needs less speed to stay in a stable orbit. Altitude and orbital speed move in opposite directions.

Drag the altitude slider and watch the orbital speed fall as the satellite moves outward, and the time dilation shrink right along with it. We write the effect as a fraction of elapsed time,

where $v$ is the satellite's speed. Negative, because the moving clock always falls behind.

General relativity: weaker gravity, a faster clock

Now for the second effect, and it points the opposite way. Einstein's general relativity says gravity doesn't just pull on things, it also slows down time itself. A clock sitting deep in a strong gravitational field, like one at sea level, ticks slower than a clock farther away where gravity is weaker.

Think of gravity as a well. Earth's surface sits near the bottom. A GPS satellite, 20,000 km up, sits much higher on the wall of that well, out where gravity's grip is looser. The higher you climb out of the well, the faster your clock runs, compared to a clock left at the bottom.

Same altitude slider as before, opposite direction of effect entirely. The formula compares gravitational potential at two heights,

Positive this time. Weaker gravity, faster clock.

Which effect wins?

Both effects are real, and at GPS altitude they don't cancel out. The gravitational speedup is bigger than the velocity slowdown, by about a factor of six. Net result: satellite clocks run fast, gaining roughly 38 microseconds every single day compared to a clock on the ground.

Thirty-eight microseconds sounds like nothing. But remember, $r = c \cdot \Delta t$. Multiply that tiny timing drift by the speed of light and it turns into real distance, fast.

Hit play and watch a simulated day pass. With the correction switched off, the position error climbs steadily, adding up to more than ten kilometers by the end of the day. Ten kilometers is the difference between "at the coffee shop" and "somewhere in the wrong neighborhood entirely."

How engineers actually fixed this

The fix is almost embarrassingly elegant. Before a GPS satellite ever launches, engineers deliberately detune its atomic clock, running it a hair slower than the standard 10.23 MHz reference rate. Slow enough on the ground that once the satellite reaches orbit and picks up its full 38 microsecond-per-day relativistic speedup, the two effects cancel and the clock ends up ticking at exactly the right rate, as seen from the ground.

This isn't a patch bolted on after the fact. It's baked into the satellite before launch, based on equations Einstein published in 1905 and 1915, decades before anyone had built a rocket that could reach orbit.

The short version

GPS solves two problems that look nothing alike. The first is geometry: three or four distance measurements, each one a circle or sphere of possibilities, narrowing down to a single intersection point. The second is relativity: satellite clocks run slow from their speed and fast from their altitude, and the two effects net out to a drift that would put you kilometers off course within a day if nobody corrected for it.

Every time your phone gets a fix, it's quietly relying on both. A little coordinate algebra from the 1600s, and a little spacetime physics from the 1900s, working together so you can find the nearest coffee shop.