Why do pure INS solutions drift over time without GNSS updates?

• A. Unmodeled accelerometer biases and random noise integrate into velocity and position, causing unbounded drift.
• B. The INS always remains accurate without updates.
• C. Atmospheric effects cause drift.
• D. The GPS satellites will drift.

Answer: (A)

Pure INS drifts because it relies on continuously integrating its internal sensor measurements to track motion, and those measurements always carry small biases and random noise. If an accelerometer has a tiny bias, that constant offset becomes a velocity error when integrated over time (roughly bias × time). Since velocity errors further integrate to position errors, the position estimate grows more and more off as time passes (roughly 0.5 × bias × time^2). Random noise adds further error in a similar accumulating way, leading to a random-walk-like drift. Without GNSS updates to provide an absolute position reference, there’s nothing to correct these accumulating errors, so the drift can become unbounded. GNSS updates supply the missing corrections, binding the INS estimates to real-world positions.

Atmospheric effects don’t cause the INS itself to drift; they affect GNSS signals, not the inertial sensors. GPS satellite motion isn’t the source of the INS drift either—the drift comes from the uncorrected biases and noise inside the INS.