A Practical Method for Evaluating Measurement System Uncertainty

This paper describes a procedure used to determine the measurement uncertainties of spherical measurement systems. This procedure requires measuring a series of unknown yet fixed points from several different instrument locations. The resulting observations are then bundled to determine the set of instrument locations that minimizes the observational discrepancies of the system. We shall consider a total measurement to be composed of three independent measurements; horizontal angle, vertical angle, and distance. The goal of this process is to derive experimentally, the individual uncertainties associated with each of these three components. By statistically processing the post-bundle discrepancies for each observation, estimates of the overall uncertainty for each component value can be obtained. We will then present how the individual component uncertainties can be used to produce realistic estimates of total coordinate uncertainty. One primary benefit of this method is that it does not attempt in any way to isolate the instrument from the measurement environment. Instead it considers the entire metrology system, including the operator, atmosphere, and targeting (SMR’s). Poor operator technique, temperature gradients and other adverse environmental effects, and poor targeting will effect the performance of the system. With this method, these effects will manifest themselves as an increase in the H, V, D uncertainty values. It is important to note that this approach will not identify the cause of poor performance, but will present an accurate and repeatable measure of how well (or poorly) the system is performing. Poor results, indicate poor system performance, not necessarily poor tracker performance. Good results, indicate good performance of the entire system.