DeepFPC: A deep unfolded network for sparse signal recovery from 1-Bit measurements with application to DOA estimation

2020 ◽  
Vol 176 ◽  
pp. 107699
Author(s):  
Peng Xiao ◽  
Bin Liao ◽  
Nikos Deligiannis
Keyword(s):  
Doa Estimation ◽  
Sparse Signal ◽  
Sparse Signal Recovery ◽  
Signal Recovery

scholarly journals Mixture Modeling based Multikernel Sparse Learning for Directional of Arrival Estimation

2019 ◽  
Vol 8 (6S) ◽  
pp. 670-675
Keyword(s):  
Expectation Maximization ◽  
Expectation Maximization Algorithm ◽  
Doa Estimation ◽  
Mixture Modeling ◽  
Sparse Signal ◽  
Sparse Signal Recovery ◽  
Sparse Learning ◽  
Signal Recovery ◽  
Uniform Array ◽  
Virtual Array

Direction of Arrival (DOA) estimation problem is defined as the problem of Sparse Signal Recovery (SSR) in researches published on the Uniform or Non Uniform array based implementations. This Paper attempts a Multikernel Sparse learning (MSL) approach with mixture modeling for the SSR problem to improve the performance parameters including the PSNR and the RMSE of the estimated sparse signal in the underdetermined condition. The Expectation Maximization algorithm is exploited to obtain the convergence in the mixture modeling MSL method. The virtual array response problem thus developed uses the mixture modeling MSL to estimate the DOA. Matlab based implementation is carried out and the results are found to be satisfactory.

scholarly journals DOA Estimation Based on Sparse Signal Recovery Utilizing Double-Threshold Sigmoid Penalty

2015 ◽  
Vol 2015 ◽  
pp. 1-8 ◽  
Author(s):  
Hanbing Wang ◽  
Hui Li ◽  
Bin Li
Keyword(s):  
Doa Estimation ◽  
Sparse Signal ◽  
Sparse Signal Recovery ◽  
Signal Recovery ◽  
Multiple Sources ◽  
Problem Model ◽  
Low Snr ◽  
Sample Covariance ◽  
Great Performance ◽  
Double Threshold

This paper proposes a new algorithm based on sparse signal recovery for estimating the direction of arrival (DOA) of multiple sources. The problem model we build is about the sample covariance matrix fitting by unknown source powers. We enhance the sparsity by the double-threshold sigmoid penalty function which can approximate thel0norm accurately. Our method can distinguish closely spaced sources and does not need the knowledge of the number of the sources. In addition, our method can also perform well in low SNR. Besides, our method can handle more sources accurately than other methods. Simulations are done to certify the great performance of the proposed method.

Sign in / Sign up

Export Citation Format

Share Document