Data Validation And Adjustment

9.1 Validation of Discrete Telemetry Data

Checking Data Flags and Sensor Identification, Validation of Discrete Data Points, Handling Invalid Data

9.2 Data Validation and Smoothing

9.3 Scalar Checking

Representative Scalars, Applications of Scalar Checking, Central Body and Horizon / Terminator Identification

9.4 Data Selection Requiring Attitude Information

Prelaunch anticipation and postlaunch analysis of erroneous data are commonly the most time-consuming aspects of attitude analysis. However, careful software design can permit automatic detection and/or correction of many types of data errors and mitigate time-consuming and costly manual data correction. For real-time operation, automatic correction or deletion of bad data is essential because of the time required for manual processing.

Bad data may be categorized in several ways:

1. According to the processing stage at which the erroneous data can be recognized and corrected, such as errors which may be identified in the telemetry processor and those which cannot be identified until an initial estimate of the attitude is obtained;

2. According to the source or cause of the erroneous data, such as transmission, hardware, software, or operator errors;

3. According to the result or manifestation of the erroneous data, such as biased output, incorrect sensor identification, or random bit errors.

The sections in this chapter are organized according to the first category. Section 9.1 describes tests which may be performed on a single frame of telemetry data and Section 9.2 describes tests appropriate to larger segments of data which may be performed in the telemetry processor or the attitude determination system and which do not require additional information such as ephemeris data or an initial estimate of the attitude. Section 9.3 describes tests requiring ephemeris information which may be done in the early stages of attitude processing before an attitude estimate is available. Section 9.4 describes tests requiring some estimate of the attitude before the tests can be conducted.

Errors which commonly occur in attitude data are summarized in Table 9-1, which is a representative sample rather than a complete list. All of the items listed have been observed in real data. Complete hardware failures, calibration errors, biases, and misalignments have not been included.

There are four general sources of error encountered in spacecraft data: data transmission, operator error, hardware or software malfunction, and "nonnominal" operating conditions. Data transmission problems, caused by a weak signal or electronic interference, increase the probability of random errors in the transmitted bit stream. No telemetry signal, or a signal below the receipt threshold, will result in data dropout, or no data. Although transmission problems may be critical to mission performance, they do not present a significant processing

Table 9-1 Representative Examples of Telemetry Data Errors (including hardware failure, calibration errors, biases, and sensor misalignments)
0 0

Post a comment