Invited Talk: Hai-tao Wang, Guilin University of Electronic Technology

Special session 6:  Advanced Technologies and Emerging Applications in Radar Signal and Image Processing

Short Bio: 
TBA

Title: A Simple Method For L₁-Regularized Least Square with Complex Variable

Abstract:
Compressed sensing or compressed sampling is widely used in radar, communication, image processing and other fields. A key problem in compressed sensing is to find an appropriate reconstruction algorithm to restore the original signal from the observed value. This problem can be cast as l1-regularized least square (LS) programs. Among the early research, to solve the l1-regularized LS, new variable vector usually should be introduced, which increase the dimension of variable vector to be solved. In this letter a simple method for l1-Regularized LS is introduced, which need not introduce new variable vector. And because a Quasi-Newton method is used, this method has lower computational complexity and memory occupation compared with traditional methods under the same signal recovery accuracy. The numerical simulation shows the effective of the method.