Efficient Maximum Likelihood Using Higher Rank Spectral Estimation
Ira J. Clarke · 2005
In practical data processing applicatiorrs, an engineer is required to develop a processing system which selects one `best' option from a span of possible interpretations of the data. A f rally optimal approach requires a search through a full range of predefine complex high rank solutions. The solutions on offer must have a sufficient number of degrees of freedom to cater ,f m complex scenarios. It is shown that a well supervised search is significantly more efficient.