Base Station Model Selection Using Machine Learning Technique for Wireless Sensor Network

The Wireless Sensor Network (WSN) is a mission-critical network technology. These networks are applied to capture necessary information from the surroundings perfectly. For delivering information to the Base Stations (BSs), the end sensor devices are responsible for sensing, transmitting, receiving and processing information in various levels of network infrastructure. Hence, it enhances the necessity to maintain efficiency, security and preservation of limited resources. In this paper, we proposed a data-driven and Machine Learning (ML) based Load Aware, Energy Efficient, and Secure Weighted Clustering Algorithm (LESWC) to deal with these qualities of Service requirements. The LESWC is combining the goodness of EE-WCA and a centralized Intrusion detection system. Additionally, some modifications have also been carried out in the data processing algorithm. First, create network clusters and then collect traffic samples on BS. The BS consists of ML algorithms to classify the traffic data and recognize secure and low-traffic routes. The proposed method offers to preserve resources in terms of buffer, energy and computational overhead. That fact is also validated through a simulation using the NS2.35 simulator. The experimental results show that LESWC is a secure, efficient and resource-preserving technique for WSN applications. The performance of LESWC has also been measured in terms of packet delivery rate, throughput, end-to-end delay and energy consumption, which shows the reliability of the proposed method.