Enhanced network lifespan in future wireless communication using machine learning based convolution neural networks

Abstract

A collection of sensor nodes called a wireless sensor network is used to track and document the physical parameters of the surrounding area. The design of network clustering approaches has a big problem when it comes to extending the lifetime of wireless sensor networks (WSNs) and improving energy usage by making sure Both the processing speed and the batteries have a lengthy lifespan. This study presents a framework for machine learning-based channel property variation tracking and learning that is based on a Convolutional Neural Network (CNN)—Long Short-Term Memory (Convolutional-LSTM) network. Our hybrid technique improves sensor connectivity and lowers power consumption, Wireless sensor network longevity is increased. These algorithms are evaluated in a wireless sensor network: Harris Hawks Optimisation (HHO), Coyote Optimisation Algorithm (COY), Support Vector Machine (SVM), and Genetic Algorithm (GA). As for nodes analysis and energy consumption, the article concludes demonstrates the CNN-LSTM technique under consideration outperforms other algorithms. The learning authentication system’s robustness and detection performance are thoroughly examined, and exhaustive simulations and testing reveal a notable improvement in the detection accuracy in time-varying scenarios.