Info

315:

o. to. ?o. 30. »0. sa m to. eo. s. is. ?s 3S si. ss. 6s. n. n.

TIHt IM niwjtis PROM siftfll une

(b) PMDICTlfr VtASUSOSSIRVtO OATA BASIS OH NOWNAL StNSOS PASAMf TIRS MO DATA FROM CtflTIR Of 1ABTWIII [F5I OMLV {NOT| FOOB QUALITY 0»

o. to. ?o. 30. »0. sa m to. eo. s. is. ?s 3S si. ss. 6s. n. n.

TIHt IM niwjtis PROM siftfll une

(b) PMDICTlfr VtASUSOSSIRVtO OATA BASIS OH NOWNAL StNSOS PASAMf TIRS MO DATA FROM CtflTIR Of 1ABTWIII [F5I OMLV {NOT| FOOB QUALITY 0»

(01 ixpailolo vimor pvot ta) shoving mgoda chamctiristics amd low amfiitudi OSCH lai(OMS oh TKf oata.

(01 ixpailolo vimor pvot ta) shoving mgoda chamctiristics amd low amfiitudi OSCH lai(OMS oh TKf oata.

Fig. 9-17. Raw Data from SMS-2 horizon sensors. (All figures are based on the same set of data.)

scan cone of the sensor enters or leaves the disk of the Earth. The effect is common to all of the SMS-2 horizon sensors.

It is clear that the data at the ends of the Earth are invalid. It is hoped that the data in the center are valid; otherwise, there is little potential for successful attitude determination. Therefore, the main question is how far into the Earth the bad data extend; i.e., what subset of the data should be used to provide the best attitude estimate. Figure 9-18 shows a plot of the spin axis declination determined from the Sun angle and the midscan rotation angle (i.e., rotation angle from the Sun to the midpoint between Earth-in and -out) from the data in Fig. 9-17(a). The data in the central region give at least somewhat consistent solutions, but the 20 to 30 frames of data at both ends are clearly part of the systematic anomaly and should* be discarded.

Data selection is performed according to the procedure described above. Both

Was this article helpful?

0 0

Post a comment