374 
measured angular speeds and the the three tilt angels roll, pitch 
and yaw : 
^ * fly * sin* 4 Jly cos X 
U> - fly • CoS X ~ J2y Sl‘*X / * 04 i 
■X 5 il 2 + tan? + a H cosx) 
The solution of this problem is possible but not the topic of 
this report. 
A block scheme demonstrates a possible system design. (fig- 1) 
Input of the sample control unit are the altitude, the airplane 
speed and the three measured values of the angular speed ¿2*, iXy 
and Sti . They are processed by the Kalman-Filter. This filter 
produces an optimal estimation of the actual angular speeds and 
predictions of the angular speeds. These predictions both are 
coupled back into the recursive filter and integrated and produce 
a prediciton of the three orientation angles. The next step is 
the derivation of the optimal sample period. This value ^ T is 
the output of the control unit. 
Of course, the reality is more complex. A sample-rate control 
using formula 3 is more realistic. It contains deviations from 
the normal speed and the normal altitude./3/ This control scheme 
shall be tested in the next future. The first step will be the 
data collection and data analysis and the off-line testing of the 
proposed algorithms. The tests will anser the question, if it is 
necessary to expand the filter and to include the altitude and 
speed data too. 
/1/ Plischke. F.: Der Einfluß von Luftfahrzeugbewegungen auf die 
Qualität von Fernerkundungsaufzeichnungen. 
Vermessungstechnik, Berlin 35 (1987) Heft 8 
/2/ Brammer K., Siffling G. : Kalman-Bucy-Filter. München: 
Oldenburg 1975 
/3/ Wirtschaftspatent WPG 01C 333 4268: Anordnung zur 
Echtzeitkorrektur fluglagebedingter Bildfehler bei Aufnahmen, 
welche von Luftfahrzeugen aus gewonnen werden. 
J.Nopirakowski, M. Scheele, Th. Terzibaschian, F. Plischke