Leveraging neural networks (RNN_LSTM) in enhancing energy efficiency and network lifetime in WSNs

Wireless Sensor Networks (WSNs) are gaining popularity due to its multiple applications in gathering and monitoring data in varying environments. However, they are associated with a few challenges such as lower network lifetime, lower resource utilization and many more. To overcome this, a hybrid deep learning approach is introduced that utilizes Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM), to enhance energy efficiency through effective cluster formation and Cluster Head (CH) selection. By doing away with the need to constantly solve intricate optimization issues inside each channel coherence time, this method drastically lowers computing complexity as compared to conventional optimization-based techniques. Numerical findings show that RNN_LSTM exhibits superior performance in terms of energy efficiency (EE) (89.03%), network stability (NS) (97.47%), network scalability (NSC) (93.76%), and QoS (91.15%) when compared to existing studies.