A Two-Dimensional Frequency Estimation Algorithm with Automatic Pairing of Cross-Dimensional Estimation Results

This paper presents a two-dimensional frequency estimation algorithm designed to automatically pair estimation results across different dimensions. The approach transforms the 2D frequency estimation problem into two generalized eigenvalue problems using matrix pencils. By computing the common eigenvectors of these matrix pencils, the algorithm simultaneously extracts eigenvalues from both matrix pencils. The estimation error is comparable to existing methods, while effectively resolving the common pairing challenge prevalent in current algorithms.