Diagnosis of array faults from far-field amplitude-only data

The diagnosis of the faulty elements of a planar array from noisy far-field power pattern data is considered in the case of "on-off" faults. The possible ambiguities of the solutions are considered both in the theoretical and practical sense and are shown to be intrinsically less relevant than in the widely studied continuous case. The probability of the occurrence of the practical ambiguities is inferred from a number of numerical examples and is shown to be negligible in all cases of interest. An effective algorithm is presented here based on an intersection set finding approach and involving the minimization of a suitable objective functional. The global minimization of the functional has been successfully performed by applying a properly modified genetic algorithm. A number of numerical examples shows the effectiveness of the approach whose computational complexity essentially increases linearly with the array size.