In this paper, an alignment algorithm is proposed based on using the outputs of inertial measurement unit (IMU).
Accurate initialization of attitude angles containing the azimuth angle is an important step to increase the accuracy of
inertial navigation system output of a moving vehicle. Two step alignment algorithm including course and fine
calculation is proposed in this paper which derives the Euler attitude angles including azimuth of vehicle with accuracy
of a few arc minutes. First a course alignment is performed based on the components of gravity acceleration and angular
velocity of the earth in body coordinate which are measured by inertial measurement unit sensors. Then the fine
alignment is performed by constructing an error model of navigation system and estimation of the attitude angles error
using a Kalman filter. The observations which are feed to the Kalman filter are the velocity components of vehicle in
navigation frame and the eastern angular velocity component of the earth which are zero. To evaluate the accuracy of
the proposed alignment algorithm, the IMU is mounted on a 3DOF table which can be aligned with 5 arc second
precision as reference attitude. Then the alignment algorithm is executed in several azimuths. The results show the
acceptable accuracy of azimuth angle.