Robust Blind Algorithm based on Oblique Projection and Worst-Case Optimization

Abstract

When adaptive arrays are applied to practical systems, the performance of the convention constant modulus algorithm degrades severely in the presence of array steering vector errors. The similar situation of performance degradation can occur even when the array steering vector is known exactly, but the training sample size is sm all. In this paper, we propose a novel doubly constrained robust constant modulus algorithm based on the worst case performance optimization and oblique projection technique. The proposed algorithm uses explicit modeling of uncertainties in the desired signal array response and in data snapshots, which provides sufficiently robustness to uncertainty in source DOA, and makes the mean output array SINR consistently close to the optimal one. The array weight vector is derived iteratively by the Lagrange multiplier approach and descent gradient technique, in which the factors can be precisely obtained at each step. A theoretical analysis for our propose d algorithm in terms of the optimal step size, convergence and array output SINR performance is presented in this paper. As compared with the linearly constrained constant modulus algorithm, our proposed robust constant modulus algorithm resolves the interference capture problem, has faster convergence speed, and enhances the array output performance under practical situations. Computer simulation results are presented to show the superiority of our proposed algorithm on output SINR enhancement and signal sampling resolution.