Sparse Blind Deconvolution Using ADMM Methods Based on Asymmetric Structured Prior for UWB Fuze.

Area of Science:

• Electrical Engineering
• Signal Processing
• Geophysics

Background:

• Ultra-wideband (UWB) fuzes require precise time delay extraction from echo signals for accurate ranging.
• Ground multipath propagation creates complex channel impulse responses (CIRs) with closely spaced components, posing a super-resolution challenge.
• Accurate CIR estimation is critical for UWB fuzes to interpret terrain and improve ranging.

Purpose of the Study:

• To develop a method for accurately estimating the channel impulse response (CIR) from single echo observations in UWB fuzes.
• To address the super-resolution problem caused by ground multipath propagation effects.
• To enhance the ranging accuracy and reliability of UWB fuzes in complex environments.

Main Methods:

• An equivalent discrete multipath model (EDMM) was established to characterize ground-based CIR.
• A sparse blind deconvolution (SBD) method was proposed using an alternating direction method of multipliers (ADMM) framework with an asymmetric structured prior (ASP).
• A periodic sparse cluster projection operator and three-variable robust decomposition were employed for super-resolution recovery and clutter separation, especially in low signal-to-noise ratio (SNR) conditions.

Main Results:

• The proposed EDMM and SBD with ASP method demonstrated high recovery accuracy for discrete multipath structures.
• The algorithm effectively separated dispersed clutter, improving performance in low SNR scenarios.
• Experimental results confirmed the algorithm's robustness and accuracy in complex, low-SNR environments.

Conclusions:

• The developed method provides a foundational offline analysis for understanding complex UWB channel characteristics.
• The approach enhances the capability of UWB fuzes to perceive terrain structures and improve ranging.
• This research guides the development of more robust and accurate real-time ranging algorithms for UWB fuzes.