On diagonal loading for minimum variance beamformers

Minimum variance beamformers are usually complemented with diagonal loading techniques in order to provide robust ness against finite sample size effects. In spite of its usefulness and popularity, some of the insights of this interesting technique have remained inconspicuous for decades. This work presents a new asymptotic performance study that will shed some light on the finite sample size behavior of this spatial filter. The asymptotic description establishes the existence of an optimum loading factor, which can be estimated from the received data.