Advances in the objective measure of comparison between antenna pattern functions

We have already highlighted the utility of a variety of objective, quantitative and robust methods of assessing similarities between antenna measurement data (see McCormick, J. et al., 2002; Gregson, S.F. et al., 2001; Gregson and McCormick, 1999). These techniques involved the extraction of interval and ordinal features from the data sets that can then be effectively compared to establish their adjacency. However, due to the volume and complexity of the data involved, a single comparison methodology is often inadequate to classify all types of data effectively. We compare and contrast several techniques for obtaining a quantitative, holistic measure of similarity between data sets and introduce a new hybrid technique. As well as conventional interval techniques, e.g. cross-correlation coefficient, two newer, more sophisticated, statistical image classification techniques are presented - an ordinal and an interval-ordinal technique. A novel hybrid categorical ordinal technique is also developed that retains the advantages of the interval-ordinal technique but removes the requirement for interpolation and facilitates the comparison of two, or higher, dimensional data sets of differing sizes. These techniques are illustrated with reference to a number of data sets that are examined, assessed and classified to obtain measures of adjacency that relate global features of the data sets. This data is derived from the output of partial scan techniques that attempt to reduce truncation errors in planar near field antenna measurements by the construction of bespoke polyhedral sampling surfaces that aim to enclose all the current sources.