Data filtering to improve the performance of the MODE algorithm for direction estimation

We propose to modify the MODE method for direction finding. The proposal uses MODE to generate a first estimate of the directions. The estimates are employed to adjust the poles and zeros of a noise reduction filter to be applied to the received data. Then the MODE algorithm is used again, now operating on the filtered data, leading to the final estimates. This procedure improves the original MODE performance and preserves the optimum asymptotic efficiency for both uncorrelated and correlated (coherent) sources.