Fast computation of iterative maximun=m likelihood estimator for continuous wave tone extraction

The short pulse length of ultra wide bandwidth (UWB) radar systems allows them to penetrate dense media such as foliage, thereby permitting battlefield surveillance of artillery and infantry. Due to their short pulse width, the UWB radars are susceptible to interference from FM and AM radio, and television stations. These narrowband interfering signals have to be suppressed to effectively operate the UWB radar system. Adaptive interference suppression techniques based on separating the eigenspaces spanned by the interference signals, and the clutter and target echoes, are proposed. Because the interference is narrowband compared to the echoes, the dimension of the vector space spanned by the interference signals is much smaller than the clutter echoes. The proposed technique suppresses interference by projecting the received data (clutter and target echoes plus interference signals) onto the orthogonal complement of the vector space spanned by the interference. The eigen-space spanned by the interference is estimated using the measured data. Computational efficiency is improved by using a proprietary algorithm that calculates the interference eigenvalues and eigenvectors more rapidly than concentional eigendecomposition routines. A computer architecture for implementing the interference suppression technique will also be developed.