Intelligent Scheduling of Sensor Nodes Using Machine Learning Techniques in Wireless Sensor Networks

Wireless Sensor Networks (WSNs) are devices that monitor physical and environmental phenomena and collect data from the surrounding environment and transmit it to the base station. WSNs have demonstrated their excellent capabilities in a variety of industries. However, sensor nodes have many limitations, such as limited battery capacity and communication capabilities. Scheduling the work of sensor nodes is essential to help conserve battery consumption and improve connectivity. This study aims to find a scheduling algorithm for sensor nodes using machine learning; an intelligent system is proposed to schedule the node in four modes automatically: active mode, safe mode, Hard-safe mode, and sleep mode, depending on the remaining power to hold the sensor. The Simulation Network-2 simulator is used to generate the data set and calculate the residual energy of nodes, and the Distance on-demand vector routing (AODV) was used as a routing protocol. Running the simulation and classification using the support vector machines algorithm proposed in this study showed the best accuracy and error rate of 100 % and 0.0 %, respectively, compared to other machine learning algorithms.