This paper presents a low-complexity robust data-dependent dimensionalityreduction based on a modified joint iterative optimization (MJIO) algorithm forreduced-rank beamforming and steering vector estimation. The proposed robustoptimization procedure jointly adjusts the parameters of a rank-reductionmatrix and an adaptive beamformer. The optimized rank-reduction matrix projectsthe received signal vector onto a subspace with lower dimension. Thebeamformer/steering vector optimization is then performed in areduced-dimension subspace. We devise efficient stochastic gradient andrecursive least-squares algorithms for implementing the proposed robust MJIOdesign. The proposed robust MJIO beamforming algorithms result in a fasterconvergence speed and an improved performance. Simulation results show that theproposed MJIO algorithms outperform some existing full-rank and reduced-rankalgorithms with a comparable complexity.
展开▼
