Exploiting time varying sparsity for underwater acoustic communication via dynamic compressed sensing

Area of Science:

• Underwater Acoustic (UWA) Communications
• Signal Processing
• Compressed Sensing (CS)

Background:

• Underwater acoustic channels exhibit multipath structures suitable for compressed sensing (CS) sparsity exploitation.
• Rapidly time-varying arrivals caused by dynamic surfaces complicate underwater acoustic channel estimation.
• Existing methods struggle with the different time-variation scales of stationary and dynamic channel components.

Purpose of the Study:

• To develop a discriminate estimation method for time-varying underwater acoustic channels.
• To model time-varying UWA channels as sparse sets with constant and time-varying supports.
• To address the challenge of channel estimation in dynamic underwater environments.

Main Methods:

• Modeling time-varying UWA channels as sparse sets with distinct support types.
• Transforming channel estimation into a dynamic compressed sensing sparse recovery problem.
• Employing a combination of Kalman filtering and compressed sensing for estimation.

Main Results:

• Numerical simulations demonstrate the proposed algorithm's superiority over existing methods.
• Field data from a shallow water experiment validate the algorithm's performance.
• The dynamic compressed sensing algorithm outperforms classic and standard CS algorithms in a decision-feedback equalizer.

Conclusions:

• The proposed dynamic compressed sensing approach effectively estimates time-varying UWA channels.
• This method offers a significant improvement for underwater acoustic communication systems.
• The algorithm successfully handles sparse components with varying time-variation scales.