A New Method of Regularization Parameter Estimation for Source Localization

摘要：

The problem of estimating the regularization parameter for source localization in sparse-regularization framework is considered in this paper. We employ the distribution about every entry of the square of the Frobenius norm of noise to obtain a larger and more appropriate regularization parameter. The paper analyzes the reason that we can not simply set it equal to the square of the Frobenius norm of noise and presents the estimation in two practical cases: one works without taking singular value decomposition (SVD) of sensor outputs; the other works after that pretreatment for large data quantity. The simulation results demonstrate that the proposed method has many advantages, including enhancing resolution, effectively suppressing spurious peaks, improving robustness to noise, as well as increasing the number of resolvable sources.

关键词： sparse representation regularization parameter estimation source localization sensor array processing

作者: Juanru Huang Mei Dong Shili Li

作者单位: National Key Lab for Radar Signal Processing, Xidian Univ., Xi’an 710071, China

会议类型: 国际会议

会议名称: 2011 IEEE CIE International Conference on Radar(2011年IEEE国际雷达会议RADAR 2011)

会议地点: 成都

会议语种:英文

页码: 1804-1808

在线出版日期: 2011-10-24（万方平台首次上网日期，不代表论文的发表时间）