A multiple time scale minimum variance beamformer

Expressions for mean output power and mean squared error are derived for the multiple time scale minimum variance beamformer in terms of adaptive degrees of freedom (DOF) and adaptation time scale when the same input data block is used to update weights and output. Analysis reveals the necessity of placing constraints on adaptation time scales to ensure the beamformer output is not ill-defined. It is shown that a signal cancellation effect arises when the same input data block is used to compute beamformer weights and form the output. The applicability of a multiple time scale beamformer when measurements consist of signals which are stationary over different time periods was demonstrated by simulation.<>