Robust signal and array processing

The importance of robustness as an attribute in engineering systems is stressed, and a new notion of robustness as applied to signal and array processing is presented. A filter is termed robust if its output is satisfactory, even when there is uncertainty regarding the input signal shape. The mathematical definitions of 'satisfactory, and 'uncertainty' as used in the formulation given here are based on hard inequality constraints. Convergent algorithms and numerical examples are presented and analyzed.