TABLE OF CONTENTS Most of the space-time processing architectures discussed in Chapters 4, 5, 6, 7, and 9 have been analysed by the author. Remarks on several processors proposed by other authors that are related to the processors discussed have been inserted into the individual chapter where appropriate. However, there are several processor concepts which do not really fit into the above scheme. A brief overview of those techniques is given in this section.
The concept of least squares prediction is well-known from the Wiener filter theory and also from adaptive coding. Transform coding is a well-known technique for data reduction in communications. Both techniques can be applied to adaptive space-time processing (Guerci and Feria [180, 181, 182, 183], Guerci et al. [184]). A predictive transform is used to reduce the signal vector space down to the required number of degrees of freedom. This results in a reduction of the number of operations required for adaptive clutter cancellation. The performance of the optimum processor very closely. When the LP processor is trained with a limited number of samples it may perform better than the SMI technique finite sample size version of optimum processor) because of the higher convergence rate.
Strong clutter discretes may lead to false alarms in adaptive clutter filter systems. Adaptation is based on averaging over a certain amount of training data so that the response of the filter is optimum for...
