Machine Learning-assisted Cluster-based Routing Protocol for Wireless Smart Grids

There are still substantial challenges to increasing energy efficiency and network longevity in wireless smart grid sensors. The current routing method uses a hierarchical WSN model to reduce the amount of energy used by WSNs. However, because of the hierarchical approach, it has significant issues with complexity, time consumption, and increased energy use. This work suggests the Machine Learning Routing (MLR) procedure for an energy-efficient WSN with a mobile sink node to address these issues. The clustering phase and the routing phase are the steps that lead to energy efficiency. K-means method starts the clustering phase by giving each node a weight value. The routing phase supports both intra- and inter-cluster routing. The Particle Swarm Optimization (ML-PSO) algorithm based on Artificial Neural Networks is suggested for intra-cluster routing. Our suggested MLR protocol maximizes network longevity by minimizing energy usage in every aspect. Promising results in terms of the number of dead nodes, average energy use, network longevity, and throughput have been shown by extensive modeling in ns-3.