Enhanced Accuracy in Underwater Communication Prediction Utilizing LSTM Over RNN for Improved Efficiency

Underwater communication systems experience obstacles like confined bandwidth, elevated energy consumption, propagation lags, end-to-end delay (E-ED), complex media accessibility control, resource allocation, and power constraints. These challenges are resolved using energy-efficient protocols, which fall into two categories: constraint-based and constraint-free. The study conducts a comprehensive assessment of constraint-free protocols in our project, considering environmental impact, data transmission rates, efficiency, energy consumption, E-ED, and propagation delay. This analysis provides insights into the pros and cons of these protocols, guiding future research in Underwater Wireless Sensor Networks (UWSNs). The study also presents routing strategies and their strengths and weaknesses, identifying areas for further investigation. In this research work, algorithms such as Long-short Term memory algorithm (LSTM), Recurrent Neural Network (RNN), Convolution Neural Network (CNN), and Artificial Neural Network (ANN), are employed and analyzed in terms of accuracy. The results demonstrate that Long-short Term memory algorithm (LSTM) outperforms better than other existing algorithms in terms of accuracy.