The Evaluation and Calibration of Range Instrumentation Systems at the Air Force Eastern Test Range

This paper gives a brief discussion of the status of work at the Air Force Eastern Test Range on the evaluation and calibration of range instrumentation systems whose accuracy specifications are a few parts per million. More emphasis will be placed on the STAFF project because of its current interest. The evaluation and calibration problems are distinct but they are intimately related. In the evaluation problem, one is asked to make estimates of the performance parameters of a range instrumentation system for purposes of understanding the system and its potential. This information is required to appraise system performance and to make recommendations for improving the system. In the calibration problem, one is asked to obtain information about the systematic errors that are stable in such a manner that this knowledge can be used to improve the accuracy of the system on future tests. In general, in the calibration of range instrumentation systems, it is desirable to suppress the so-called systematic or bias errors to a level below the random errors. Solutions to both the evaluation problem and the calibration problem are required for an orderly operation and development of a research and development missile and space test range, such as the Air Force Eastern Test Range. There is no accuracy standard available in the conventional sense for use in the evaluation and calibration of some instrumentation at the Air Force Eastern Test Range, since the systems under study are the most accurate systems known. This has prompted the Range to turn to statistical techniques involving regression techniques, non-linear estimation theory and time series. The ETR does not feel that the evaluation and calibration problem can be solved entirely by statistical techniques. It is necessary to tie the hardware (including special tests) intimately to the software. In general, the error model parameters studied must have physical or hardware significance, such as refraction, drift of oscillator and timing errors. Results found by statistical methods must be verified and understood in terms of the hardware performance and the physical situation. The procedure involves checking results obtained by statistical methods by special subsystem or other testing to confirm them. Once the errors are understood in terms of the hardware or in physical terms, the first approach is to improve the hardware or the understanding of nature to reduce the error to acceptable standards. Only when this cannot be done, with reasonable effort, can one be satisfied to estimate and attempt to correct the effect of the error by software techniques. Even in the case where software methods are predominant, special subsystem tests and improvement in the hardware is of fundamental importance since it may reduce the a priori variances of the error and possibly reduce the number of parameters to be estimated, thereby increasing the degrees of freedom and the power and the accuracy of the results