
Why Navigators Never Stop Correcting Course
Every position estimate drifts until another observation pulls it back

Every position estimate drifts until another observation pulls it back
AI-assisted edition · Educational review score 96%
Every position estimate drifts until another observation pulls it back
Created by Bob · AI-assisted and reviewed before publicationDead reckoning begins with a known or estimated position, then advances it using measured direction, speed, and elapsed time. A walker counts steps and heading; a ship combines course and speed; an inertial system integrates acceleration and rotation.
Between outside observations, this prediction is often the only available position. It's useful precisely because it continues through darkness, tunnels, or signal loss. But the new point inherits every error in the old point and adds errors from the motion measurements.

Random noise may partly average out, but a persistent bias pushes the estimate in one direction. A compass offset bends the route; a speed error stretches it; an accelerometer bias becomes velocity error and then position error when integrated over time.
Wind, current, wheel slip, and human steering add unmeasured motion. The navigator may know the estimate is becoming less trustworthy even without knowing the exact direction of the error. Uncertainty expands with time and movement since the last reliable fix.

An external observation; landmark bearing, radio range, GPS position, map match, or star sight; compares the prediction with evidence tied to the world. A navigation filter weighs both according to their uncertainty, moves the estimate toward agreement, and reduces the range of plausible positions.
The correction isn't perfect because the observation also has noise and possible bias. Navigation repeats this cycle: predict motion, observe, compare, correct, and carry the revised uncertainty forward. A route is maintained, not just known once.

These references were used to check the important factual claims in this edition.