Machine Learning-based Clustered Data Dissemination Protocol for Mobile Wireless Sensor Networks

Mobile Wireless Sensor Networks (MWSNs) are widely operated in dynamic environments, including smart cities, environmental monitoring, and disaster management. However, mobility-induced topology variations introduce significant challenges in clustering and data dissemination. To address these challenges, an Energy-Efficient Machine Learning-Optimized Cluster-based Data Dissemination (EMCDD) Protocol was proposed. This protocol integrates a hybrid model of Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) for optimal cluster formation and CH selection, while Support Vector Machine (SVM)-based classification ensures robust CH election. The data transmission is optimized using a deep Q-learning-based adaptive scheduling mechanism, which dynamically allocates time slots based on node mobility and connection duration. Simulation results demonstrate that EMCDD outperforms HDDP and ADDP protocols in terms of throughput, total energy consumption, and end-to-end delay under varying node mobility and traffic conditions.