IMU In-Motion Alignment Without Benefit of Attitude Initialization

Establishment of local-level wander azimuth navigation and body reference frames for an inertial navigation system (INS) based on outputs of an inertial measurement unit (IMU) without the use of initial attitude information is addressed. These frames are established using a Kalman filter algorithm implemented with an INS system error model formulated for large heading errors and using position measurement updates. Only position and velocity in a geographic reference frame are used for initialization of the INS navigation equations. This data is available from GPS receivers; however, in this paper, an aircraft navigation system's data is used instead to demonstrate the alignment performance using data from actual flight tests.