Fast design of optimal array filters

Array processing is applied to signals arising in seismology, sonar, radar, astronomy and remote sensing. Optimal filtering of the sensor signals utilises the spatial diversity of the array receivers so that a desired signal embedded in coherent noise signals and random noise may be extracted. This paper provides an analytic solution for the optimal array filter for the case where the desired wide-band signal is all-pass filtered and a single coherent interference is nulled. The filter coefficients are computed without matrix inversion and the solutions take the form of rational polynomials with a common denominator term that can be interpreted in terms of a spatial sampling requirement. As the OAF has an all-pass response with minimum delay for the desired output signal, the array filter structure is seen to be a standard signal summer followed by a post-interference canceller (PIC)