On the performance of the alternating maximum likelihood notching algorithm

Simulation results are presented to demonstrate that the alternating maximum likelihood notching algorithm (AMNA) outperforms the iterative quadratic maximum likelihood (IQML) algorithm for low signal-to-noise-ratio (SNR) cases. Moreover, it is shown that further improvement of the AMNA can be accomplished by using an elaborate two-dimensional slant search to adjust two bearing estimates simultaneously. This not only improves the accuracy but also results in faster convergence speed and higher resolution, especially for the cases of closely spaced sources. In addition, the algorithm is very suitable for hardware implementation through the concurrent Schur recursions and the associated pipelined combiner-lattice structure, thus making this improved algorithm more competitive in practical applications