J4 ›› 2014, Vol. 41 ›› Issue (5): 30-35.doi: 10.3969/j.issn.1001-2400.2014.05.006

• Original Articles • Previous Articles     Next Articles

Improved low-rank recovery method for sparsely sampling data in array signal processing

YANG Dong1;LIAO Guisheng1;ZHU Shengqi1;WANG Kai2   

  1. (1. National Key Lab. of Radar Signal Processing, Xidian Univ., Xi'an  710071, China;
    2. Unit 91251, PLA, Qinhuangdao  066102, China)
  • Received:2013-05-29 Online:2014-10-20 Published:2014-11-27
  • Contact: YANG Dong E-mail:yangdongxd@gmail.com

Abstract:

Matrix Completion (MC) theory can recover the under-sampled data in the array signal processing, further estimating the direction of arrival (DOA) as the fully sampled data does. However, it is required that the data should be under-sampled randomly in different snapshots which satisfy the randomness of MC theory. When some sensors are unsampled or broken in the whole observing time, the previous method would fail. To address this problem, a new processing method is proposed in this paper. The inner relationship among sensors is used, and then we reshape the signal vector in a single snapshort into an equivalent low-rank matrix, which can be recovered effectively by minimizing the nuclear norm. Simulation results validate the effectiveness of the proposed method. Meanwhile, the method can lower the noie power, and improve the performance of the DOA.

Key words: array signal processing, matrix completion, low rank matrix, direction of arrival