A Feature Pattern Tracking Method Based on Feature Vector Correlation

Abstract

This invention relates to a feature mode tracking method based on eigenvector correlation, belonging to the field of antenna technology, and includes two parts: preprocessing and post-processing. The preprocessing filtering method, targeting different radiation structures, retains only modes that significantly affect antenna performance. The post-processing sequentially addresses mode switching and mode degeneracy issues. In the mode switching problem stage, the tracking curve intersections are first identified, and a variance criterion is applied to determine if mode switching exists at the intersections. Then, eigenvector correlation is used for re-matching. In the mode degeneracy problem stage, degenerate mode pairs are first identified, and an energy criterion is applied for judgment. Then, degenerate mode pairs are corrected, and finally, the tracking curve is optimized to achieve a smooth tracking curve. Compared to traditional methods, this invention reduces the number of modes in the preprocessing stage, saving subsequent computation time, and classifies and handles mode tracking errors in the post-processing stage, ensuring the accuracy of the tracking results.

Images