Efficient Resource Optimization Techniques to Enhance Routing and Data Delivery in Wireless Sensor Networks

Abstract

The usage of wireless sensor network (WSN) has been dominant since few decades driving several applications in consumer, industrial, medical services and etc. Generally, a WSN consists of sensor nodes, sensor cluster, and cluster heads, and is operated using specific wireless communication protocols. There are so many challenges found while implementing such a network. It includes improving network performance in the context of energy efficiency, data aggregation, security attacks, channel modeling, scheduling and etc. This research addresses the first three aforementioned parameters which are modeled and analyzed using few new techniques. Firstly, a Fully Connected Energy Efficient Clustering (FCEEC) algorithm is proposed to achieve full connection and efficient packet distribution from a cluster head (CH) to Base station (BS) nodes. This technique utilizes the electro static discharge algorithm (ESDA) to construct a network with shortest path routing from sensor nodes (SNs) to cluster head. As these sensor networks are battery power operated, energy conservation plays a prime factor. Thus, smooth clustering schemes are drafted to preserve node energy to extend the network life time. Secondly a machine learning based data aggregation model is proposed to mitigate redundant data process and transmission. The model employs a Machine Learning based Extreme Learning Machine with Adaptive Kalman Filter newlinevi newline(MLELMAKF) for data aggregation in which redundant data is predicted and eliminated before transmission. Thirdly an Intrusion Detection System (IDS) is proposed using a new Network Intrusion Detection - Machine Learning Classifier based Intrusion Detector for Communication (NID_MLCIDC). The goal of this model is to use machine learning (ML) classifier to classify various security attacks in WSNs and then to predict future attacks with reliable values of accuracy. newlineThe performance metrics are presented in terms of accuracy, precision, F-Score, false alarm, energy efficiency, network life time, and etc. There have been