Hybrid Positioning and Sensor Integration - Global Navigation Satellite Systems: Signal,Theory and Applications - page 345

selected, and covariance must be initialized in order to continue Kalman filtering circulation
in response to the GPS data validation. The accuracy of the integration depends on the
accuracy of the referenced GPS. In this case, it is approximately 30 cm.

Delta angle IMU attitude Delta velocity

Initial
attitude

Euler
matrix
generation

Coordinate
transformation

Quaternion
update

Ve l o c i t y g e n e r a t i o n
coriolis compensation
gravity compensation

Torque rate
generation

IMU velocity

Initial position
initial wander
azimuth

Relative rate
generation

Latitude /
longitude
wander azimuth
generation

Earth rate
generation

Fig. 3. Pure navigation block diagram expressed roughly step by step

Inertial estimate
and its covariance New measurement

GPS on/off

Project ahead

Update covariance
matrix

Measurement
matrix calculation

Update filter
estimate

Kalman gain
generation

Fig. 4. Kalman filter circulation

This research adopts the Kalman filter and the following steps are included, as shown in
Figure 4.

1. Initialization of covariance value and first estimation of each variable

2. Inspection of GPS validity and selection of measurements equation