METHOD FOR DETECTING FAULTY MEASUREMENT EVENTS AND COMPUTER PROGRAM PRODUCT

The invention relates to a method for detecting faulty measurement events and a computer program product with means for performing the method according to the invention. [0002] In this method, a monitoring period (T) is subdivided into many short, equal-length time intervals (t), in which characteristic variables are determined and the time intervals (t) are subjected to categorization into "ordinary conditions" or "unusual conditions". The characteristic variables are either the number of arcs per time period n(t) (variant 1) or the mean distance s(t) traveled between two arcs (variant 2). After comparison of the characteristic variables with a threshold value determined from the mean value of all characteristic variables (n, s) of the monitoring period (T) and, if necessary, an appropriate dispersion measure (e.g. standard deviation) σ or σ, the time intervals (t) are classified as time intervals with unusual conditions or with faulty measurement events if the corresponding characteristic variable of the time interval (t) is greater than the threshold value.