A near-field global-correlation method is proposed for accurate tracking of a larger number of dense characteristic modes in a large bandwidth. By using characteristic near-field patterns instead of conventional eigen-vectors or far-field patterns, the information stored in the rapid change of surface currents and the near-field evanescent waves is fully leveraged, thereby enriching the information of the mode database for accurate mode-tracking at the data level. To make full use of the enriched mode information, a global-correlation algorithm is proposed to correlate both the rows and columns of the correlation matrix, rather than conventional local correlation that concerns only the rows or the columns. The proposed method is validated by four examples of different complexities, including a rectangular plate, a bow-tie plate, a ground-backed fractal patch, and a ground-backed metasurface. As a result, both the bandwidth and accuracy of multi-mode tracking are significantly improved, as characterized by the bandwidth and the number of modes tracked without mode-swapping or mode-missing.
You et al. (Sun,) studied this question.