1: \begin{abstract}

2: 

3: %In dynamic propagation environments with large sensor arrays,

4: %beamforming algorithms may suffer from strong interference, steering

5: %vector mismatches, a low convergence speed and a high computational

6: %complexity. Reduced-rank signal processing techniques provide a way

7: %to address the problems mentioned above.

8: This paper presents a low-complexity robust data-dependent

9: dimensionality reduction based on a modified joint iterative

10: optimization (MJIO) algorithm for reduced-rank beamforming and

11: steering vector estimation. The proposed robust optimization

12: procedure jointly adjusts the parameters of a rank-reduction matrix

13: and an adaptive beamformer. The optimized rank-reduction matrix

14: projects the received signal vector onto a subspace with lower

15: dimension. The beamformer/steering vector optimization is then

16: performed in a reduced-dimension subspace. We devise efficient

17: stochastic gradient and recursive least-squares algorithms for

18: implementing the proposed robust MJIO design. The proposed robust

19: MJIO beamforming algorithms result in a faster convergence speed and

20: an improved performance. Simulation results show that the proposed

21: MJIO algorithms outperform some existing full-rank and reduced-rank

22: algorithms with a comparable complexity.

23: 

24: 

25: 

26: \end{abstract}

27: