Employing feedforward backpropagated neural network for Doppler scale estimation in underwater acoustic CP-OFDM...

Area of Science:

• Electrical Engineering
• Signal Processing
• Underwater Communications

Background:

• Orthogonal Frequency Division Multiplexing (OFDM) is a key technology for underwater acoustic (UWA) communication.
• UWA channels present significant challenges, including large multipath and severe Doppler effects, which degrade communication performance.
• Accurate Doppler scale estimation is crucial for mitigating these effects in UWA OFDM systems.

Purpose of the Study:

• To propose a novel feedforward backpropagated neural network (FBNN) implementation for Doppler scale estimation in UWA cyclic-prefix (CP) OFDM systems.
• To evaluate the performance of the proposed FBNN using different backpropagated training algorithm variants.
• To compare the FBNN approach with conventional methods for Doppler estimation.

Main Methods:

• A two-layered input-output feedforward neural network architecture was designed.
• Three training algorithm variants were employed: Fletcher-Reeves Conjugate Gradient (CGF), Polak-Ribiére Conjugate Gradient (CGP), and Conjugate Gradient with Powell/Beale Restarts (CGB).
• Root Mean Square Error (RMSE) was used to assess performance under various multipath and signal-to-noise ratio (SNR) conditions.

Main Results:

• The proposed FBNN effectively estimates the Doppler scale factor by integrating neural network capabilities with the precision of conjugate gradient training algorithms.
• Performance was evaluated across diverse channel conditions, demonstrating the robustness of the FBNN approach.
• Comparative analysis showed the FBNN-based methods achieved competitive or superior performance against benchmark conventional techniques.

Conclusions:

• The FBNN implementation offers a promising solution for accurate Doppler scale estimation in UWA OFDM communication.
• The combination of neural computation and advanced training algorithms enhances resilience to multipath and Doppler effects.
• This work contributes to improving the reliability and efficiency of underwater acoustic communication systems.