Fast estimation of sparse doubly spread acoustic channels

Area of Science:

• Signal Processing
• Underwater Acoustics
• Optimization Theory

Background:

• Underwater acoustic channels are time-varying and exhibit complex propagation characteristics.
• Accurate channel estimation is crucial for reliable underwater communication.
• Existing methods for complex-valued optimization can be computationally intensive and may alter data structures.

Purpose of the Study:

• To develop a computationally efficient method for estimating doubly spread underwater acoustic channels.
• To address the limitations of conventional complex-valued optimization techniques.
• To leverage sparsity in the delay-Doppler domain for improved channel estimation.

Main Methods:

• A Fast Projected Gradient Method (FPGM) is proposed for sparse channel estimation.
• The method formulates the problem as a complex-valued convex optimization using an L1-norm constraint.
• FPGM directly handles complex variables, avoiding dimension increase, and exploits block Toeplitz-like structures for efficiency.

Main Results:

• The proposed FPGM achieves a computational complexity of O(LNlogN).
• Simulations demonstrate the accuracy, efficiency, and robustness of FPGM to parameter selection.
• The algorithm is orders-of-magnitude faster than standard convex optimization methods.

Conclusions:

• FPGM provides an accurate and efficient solution for estimating complex-valued, time-varying underwater acoustic channels.
• The method's direct handling of complex variables and exploitation of channel structure offer significant advantages.
• Experimental data processing confirms the practical performance of the proposed FPGM algorithm.