Statistical spectral analysis: a nonprobabilistic theory
Statistical spectral analysis: a nonprobabilistic theory
| Hi-index | 0.00 |
A new cost function is proposed for adaptive beamforming that exploits the property of cyclostationary signals to generate spectral lines by use of a particular class of nonlinear transformations. Analysis results are presented which demonstrate that the beamformer can extract the desired signal using only the information from the spectral lines. It is also shown that interference can be eliminated even when it exhibits the same cyclostationary properties as the desired signal. This new approach is particularly useful for continuous adaptive implementation since the optimal weights can be computed using a simple stochastic gradient algorithm.