The same principles are used in the RT. Position and velocity are compensated directly,
but other measurements like accelerometer bias, have no direct measurements. The
Kalman filter tunes these so the GNSS measurements and the inertial measurements
match each other as closely as possible.
The Kalman filter in the RT has 24 states. These are position error (north, east, down);
velocity error (north, east, down); heading error; pitch error; roll error; gyro bias (x, y, z);
gyro scale factor (x, y, z); accelerometer bias (x, y, z); GNSS antenna position (x, y, z);
GNSS antennas orientation (heading, pitch) and vehicle mounting angle (for advanced
slip).
The errors are applied smoothly to the states. For example, if the Kalman filter wants to
correct a position error of five centimetres in the north direction then this is applied
slowly, rather than jumping directly to the new position. This helps applications that use
the RT for control since any differential terms in the control algorithm do not have large
step changes in them.
To Next Page
Loading...
Need help?
General
