On the robustness of several array processing algorithms

Robustness in array signal processing has several different meanings. Most generally, it refers to algorithms with performance that is insensitive to incorrect assumptions. These assumptions can take different forms, leading to different notions of robustness. This dissertation consists of three studies of array signal processing algorithms, with respect to different notions of robustness. The first study is an investigation into the numerical robustness of a beamforming algorithm. Numerical robustness refers to the ability of an algorithm to perform well when implemented on finite-precision hardware. The second study analyzes the performance of a blind signal separation algorithm operating in an environment significantly different from that which it was designed for. The last study investigates the effect of incorrect modelling of the probability density function of additive noise on a array processing signal detection algorithm.