Interference suppression via orthogonal projections: a performance analysis

Several recent studies indicate the promise of subspace separation principles when applied to adaptive jammer suppression in phased array antennas. This paper theoretically analyzes the performance of a subspace separation technique based on orthogonal projections (OP) for adaptively suppressing interference in phased arrays; the theoretical performance predictions are validated using computer simulations. This analysis holds for the case when it is possible to differentiate between the vector spaces spanned by jammers and additive noise. The performance parameters used are (a) the average residual interference (jammer plus noise) power at the output of the adapted array as a function of the number of jammer snapshots used for calculating the weight vector, and (b) the similarity of the adapted array pattern to the design pattern away from the jammer locations. The performance of the OP-based subspace separation technique is compared with the sample matrix inversion (SMI) algorithm. It is shown that the weight vector calculated using OP converges more quickly to the optimal solution (infinite number of interference snapshots) than the SMI weight vector. Further, in contrast to the SMI adapted pattern, which exhibits large sidelobe levels away from the jammer locations, the OP adapted pattern closely follows the design pattern both in the mainbeam and in the sidelobe region away from the jammer locations