9.4 Data Selection Requiring Attitude Information

9.4 Data Selection Requiring Attitude Information

James R. fVertz

Previous sections have described data selection or validation procedures which do not require any knowledge of the attitude and, therefore, can be performed at an early stage of attitude processing. In contrast, data selection requiring attitude information must be a part of the attitude determination process itself and may require an iterative procedure to determine if the data selection process is consistent with the computed attitude and to reselect the data if it is not.

The most straightforward data selection of this type is that which requires only an a priori attitude estimate, that is, an attitude estimated or assumed before any processing is done. In practice, some a priori knowledge is usually available and this is generally sufficient to resolve quadrant ambiguities or to choose the correct attitude solution from the two possible solutions generated by intersecting cones. The latter procedure is described in more detail in Section 11.2. For spacecraft using automatic control, the intended or null attitude may be used to validate data used for a definitive attitude solution; however, any such test may also effectively hide a failure of the control system, since data that is inconsistent with the intended attitude would automatically be rejected.

When data selection requires attitude information, it may become the most complex and time-consuming aspect of the attitude determination process. For example, in Fig. 9-2, an attitude estimate and manual data editing were required to determine which of the two groupings of data at the top of the figure was valid and which was anomalous.

On RAE-2 a horizon sensor of a new design (the panoramic attitude sensor described in Section 6.2.2) was flown. During the translunar portion of the flight, much of the data was spurious. Figures 9-16(a) and 9-16(b) illustrate displays which were used to manually distinguish valid lunar sightings from spurious data due to the Sun, spacecraft reflections, or noise. The ordinate of Fig. 9-16(a) is the angle, y, from the spin axis to the scanner line of sight and the abscissa is the rotation angle, from the Sun to a sighting event. The panoramic scanner is a variable mounting angle sensor; for RAE-2 the angle y changed by 0.707 deg approximately every 15 sec. The crosses in Fig. 9-16(a) mark observed light to dark (LOS) or dark to light (AOS) transitions and the ovals mark the expected location of solar and lunar data for the a priori attitude 5 hours before insertion into lunar orbit. The data observed near $ = 0 deg or 180 deg and 60 deg>y>0 deg are clearly spurious and are believed to have been caused by reflected sunlight [Werking, 1974]. The relatively smalt amount of data near i> = 65 deg and y = 30 deg are valid AOS or LOS events from the lunar horizon or terminator. The expanded view of these data in Fig. 9-16(b) shows that most of the valid data were LOSs at the terminator. Closer



160. 170. 160. 150. >>40. 130. 120. 110. 100. 90. 80. 70. 60. 50. HO. 30. JO. 10. 0.

0 0

Post a comment