How can the accuracy be greater or less than the specified accuracy at different points in the motion box?
A. Ascension specifies the tracker’s accuracy using the Root-Mean-Square (RMS) error statistic. It is computed by moving the sensor to hundreds of locations throughout the motion box and recording the error at each location. The RMS value is then computed using all the error data throughout the motion box. RMS = square root of (sum of all errors squared / number of locations). Statistically speaking, if the errors are normally distributed then 68% of all errors fall within the band of +/- RMS and 99.7% of the errors fall within the band of +/- 3 x RMS. For some Ascension trackers the RMS accuracy is specified as 0.07 inch or 1.8 mm. Other products are specified as 0.04 inches or 1.0 mm. A sensor placed at any one location in the motion box will read only the peak error only at that location, not the RMS error. As a result, some errors may be larger and some smaller than the specified accuracy. For example, if the sensor were located at the peak positive error then you would measure an
