A low-complexity implementation of adaptive Wiener filters

Wiener filters of arbitrary order can be implemented by a nested structure of one-dimensional filters. The multistage processor is simple to implement, and covariance matrix estimation and inversion are not required to calculate the filter weights. The nested structure provides a new framework for designing and implementing reduced complexity filters. When applied to a typical array processing problem, it provides close to optimum performance, even when its degrees of freedom are less than the rank of the interference subspace.