Why is quaternion normalization applied after attitude updates?

• A. To maintain the unit-norm constraint and prevent attitude drift.
• B. To align the attitude with the gravity vector.
• C. To simplify the Kalman gain computation.
• D. To reset attitude to the initial value.

Answer: (A)

Quaternions representing rotations must have unit length, so after attitude updates the computed quaternion can drift away from 1 due to numerical errors and finite-precision arithmetic. Normalizing brings it back onto the surface of the unit sphere (the set of unit quaternions), preserving the direction of the estimated rotation while enforcing the correct magnitude. This keeps the attitude representation valid and prevents artificial attitude drift that would distort the orientation over time. It isn’t about aligning with gravity, nor about simplifying Kalman gains, nor about resetting to the initial value—it’s a corrective step that maintains a proper, stable rotation quaternion.