An indirect measurement is a measurement in which the value of an unknown quantity (a measurand) is estimated using measurements of other quantities (arguments) related to the measurand by a known functional dependency. This paper addresses the basic problems inherent in indirect measurements. It demonstrates that the traditional approach to estimating a measurand is biased when the dependency between the measurand and any relevant arguments is non-linear. In addition, the calculation of variance of this estimation of a measurand requires the calculation of a correlation coefficient. This calculation entails many well known problems. This paper describes an alternative technique: the Method of Reduction which is free from the above mentioned deficiencies. The Method of Reduction produces a non-biased estimate of a measurand for both linear and non-linear indirect measurements. Variance of estimation of a measurand is easy to obtain by this method and does not require a correlation coefficient at all. The Method of Reduction requires that all arguments be measured under specific plan. Because of its simple, direct quality, the Method of Reduction is proposed as the basic method used for the determination of indirect measurements. In addition to the Method of Reduction, this paper discusses the problems of estimation of errors and the uncertainty of indirect measurement results.
展开▼
