Figure 3. Velocity with Kalman filter 
880 
’r ¡¿g 
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B5. Beijing 2008 
The Kalman filter has been applied, and the results of position, 
velocity and acceleration obtained without and with Kalman 
filter, have been compared. 
Figures 1 and 2 show the acceleration without Kalman filter and 
with Kalman filter, for 9000 and for 500 acquisitions. Figures 3 
and 4 show velocity and position obtained with Kalman filter. 
Figure 5 shows the position without Kalman filter. 
— x acceleration without Kalman filter (m/sec A 2) 
— x acceleration with Kaiman filter (m/sec A 2) 
0,08 -r — i— , 
O O a |Q O O o I o 
-S H R -3- 1 -a § Pi 8- 
sample 
Figure 1. acceleration along x axis without and with Kalman 
filter 
Figure 2. x acceleration for 500 acquisitions 
x velocity with Kalman filter (m/sec) 
Figure 4. Position with Kalman filter 
Figure 5. Position without Kalman filter 
The above drawings show the Kalman filter effect on the 
accelerations and the decreasing drift of the position. 
4. THE ACCELEROMETER CALIBRATION 
The described multi position calibration method has been used 
for determination of bias, scale and misalignment factors of the 
ADIS 16350 IMU. 
The procedure has been applied only for the tri-axial 
accelerometers. 
The IMU has been placed in 39 different positions: the angle 
between two successive positions is about 45°. The orientations 
have been obtained by positioning the sensor on a rotating 
instrument (Figure 6). The precise rotation is unknown, 
however, in multi position calibration method, the precise 
attitude of sensors is not required; only the independence of 
equations for the rapid convergence of results is needed.