An Energy-Efficient Clustering Algorithm for Maximizing Lifetime of Wireless Sensor Networks using Machine Learning

Wireless sensor nodes come with small-sized batteries that provide the energy required to do all kinds of activities. Unnecessary usage of the radio, especially during idle listening, drains a lot of energy in Wireless Sensor Networks (WSNs). The proposed model, Energy-Efficient Clustering Algorithm (EECA), prolongs the WSN lifetime by minimizing energy consumption in sensor nodes. In EECA, the target area is considered to be a collection of small regions. The proposed model uses Artificial Neural Network (ANN) to select one node in each region as the cluster head (CH). The sensor nodes with a predefined minimum energy level qualify for the CH selection process. ANN calculates the scores of these nodes based on four parameters - residual energy, number of events detected, distance to the base station, and number of neighbours. The sensor node with the highest score is selected as the CH in its region. A limit called maximum cluster size is also defined to avoid the formation of huge clusters. Only the sensor nodes located close to an event inform the CH about the event. This rule ensures that redundant data is not transmitted to CHs. In the proposed model, a CH checks the medium for a very short duration at the beginning of a slot for incoming transmissions. If the CH does not receive any signal within this duration, it turns its radio OFF. This rule minimizes idle listening in CHs. EECA is compared with some existing medium access control protocols to find out its efficiency. The experimental results show that EECA saves more energy in sensor nodes than in other models.