Sparsity enhancement post-nonlinear blind deconvolution method and its application to aluminum honeycomb panel cabin structure

In this paper, we propose a method for post-nonlinear blind source separation. The method divides the separation process of post-nonlinear mixed signals into two independent stages: the nonlinear compensation stage and the linear blind source separation stage. The nonlinear compensation stage is achieved by taking sparsity enhancement as the optimization objective. The L1-norm is taken as the objective function and is combined with the fast iteration based on the gradient descent method to realize the fast nonlinear compensation of the mixed signals. In the stage of linear blind source separation, the blind deconvolution algorithm with reference signals is used to process the compensated signals to realize the separation of the source signals. The separation performance of the method is verified by simulation, and the superiority of the method is tested by comparison. The proposed method is also investigated by the excitation experiment of the aluminum honeycomb panel cabin structure, which simulates the satellite structure. (paper)