Sparse regularization for reconstructing transient sources with time domain nearfield acoustical holography

Area of Science:

• Acoustics
• Signal Processing
• Numerical Methods

Background:

• Time domain reconstruction of transient acoustic fields is crucial for understanding acoustic phenomena.
• Standard Tikhonov regularization methods often introduce causal errors in transient sound field reconstruction.
• Impulse noise and other transient sources present unique challenges for accurate acoustic field reconstruction.

Purpose of the Study:

• To apply the ℓ₁-norm sparse regularization method for improved time domain reconstruction of transient acoustic fields.
• To compare the effectiveness of sparse regularization against standard Tikhonov regularization.
• To investigate the performance of the Pareto frontier curve for optimizing sparse regularization parameters.

Main Methods:

• Application of ℓ₁-norm sparse regularization to transient acoustic field reconstruction.
• Comparison of reconstruction results using Tikhonov and sparse regularization with a transient baffled piston model.
• Analysis of the Pareto frontier curve for selecting optimal regularization parameters.
• Validation with experimental data from an impacted plate.

Main Results:

• Sparse regularization significantly reduces causal errors in transient sound field reconstruction.
• The global root-mean-square (RMS) error was reduced by up to a factor of 3 compared to Tikhonov regularization.
• The Pareto frontier curve accurately predicts optimal sparse regularization parameters, particularly at lower noise levels.
• Successful application to experimental data demonstrates practical utility.

Conclusions:

• ℓ₁-norm sparse regularization offers a superior method for reconstructing transient acoustic fields.
• This technique effectively minimizes causal errors and reduces overall reconstruction error.
• Sparse regularization shows significant potential for accurate transient sound field analysis in both simulations and experiments.