Procreation of Energy Efficient Hierarchical Routing Protocols for Wireless Sensor Networks

Abstract

Wireless Sensors Network (WSNs) is a self-configured and infrastructure-less network. In the past few decades, it has been explored by many researchers for energy-efficient communication. WSNs are utilized in several real-life applications such as monitoring environmental conditions, health, security surveillance, target tracking, etc. Typically, a WSN is a collection of tiny sensor nodes powered by diminutive power sources. These sensor nodes sense phenomenal data continuously and report it to the base station using single or multi-hop communication. Hence, the network lifetime decreases continuously as the battery drains quickly while data sensing, its' processing and remote transmission. Usually, these wireless nodes are deployed in potentially harsh physical conditions, and are not feasible to change the batteries or even replenish them. Researchers have suggested various approaches to reduce the energy consumption in WSNs, such as radio optimization, reduction in data to be transmitted, sensor sleep/awake mechanism, energy harvesting, wireless charging, and energy-efficient routing. Transmission of data is supposed to be the primary source of energy depletion in WSN. Thus, energy-efficient routing and routing protocol design play an essential role in enhancing the WSN's network lifetime. The energy consumption rate of traditional WSN routing protocol is high and possesses relatively small network lifetime. They suffer from hotspot energy-hole related problems. Therefore, various types of routing protocols, including location-aware, data-centric and hierarchical, have been proposed. Location-aware protocols take benefit of location awareness of sensor nodes. Awareness of Geographic location may help estimate energy consumption based on the calculated distance among sensor nodes. Moreover, by using location information of sensor nodes, enquiry for information may be communicated to the specific area, and thus the number of transmissions can be reduced significantly. Here, the emplacement of geographical localization devices and their credibility can be significant defiance. Moreover, the sensors' energy consumption is sheer proportional to the distance from the target node to which data is to be communicated. It implies that, with the increase in the data transmission distance, the energy consumption increases exponentially. Data-centric protocols perform data fusion to remove redundancy and transmission. These protocols take advantage of attribute-cantered naming conventions rather than addresses. The sink node remits the desired query to particularly identified areas and expects a response from distributed sensor nodes in that area. The major limitation of data-centric routing protocols is its' scalability. It is due to the involvement of volumetric sensor nodes in the WSN. Hierarchical or clustering-based routing protocols divide the WSN into clusters to achieve energy-efficient communication. Each cluster has a Cluster Head (CH) and is amenable to data collection, aggregation, and forwarding it to the next CH or the BS in single/multiple hops. The challenges with hierarchical routing protocols lie in the appropriate division of the network into clusters, suitable CH nomination, and data communication in an energy-preserving way. The hybrid optimization-based CH election primarily depends on some key parameters such as energy, distance, and delay. Here the parameters like connectivity and coverage range have given less importance. Another aspect of such protocols is selecting an optimal path from the sensor node to the sink node. The leading cause of such a problem is the existence of multiple multi-hop paths. The research work in this dissertation presents three protocols for CH selection and energy optimization. The two protocols are nature inspired and are energy efficient protocols namely FMCB-ER and MCH-EOR and third one is an improvement to the pre-existing protocol called Node-Rank-LEACH protocol and is known as LBNR-LEACH. These protocols enable effective CH selection and energy efficient path selection for communication. The FMCB-ER (Fuzzy Multi-Criteria Clustering and Bio-inspired Energy-efficient Routing) is a hierarchical-WSN protocol, MCH-EOR is a "Multi-Objective Cluster Head-based Energy-aware Optimized Routing" protocol, and LBNR-LEACH considers the network load into consideration along with the link quality to select best CH. FMCB-ER is structured on clustering using a grid approach. A grid has been used to divide the network into robust and distributed clusters for efficient data delivery. The F-AHP (Fuzzy-Analytic-Hierarchy-Process) and TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution) methods are used to select an optimum CH based on three criteria, namely energy status, node position, and impact on QoS and their six sub-criteria. Each criterion is assigned some weights using AHP, and TOPSIS ranks the CHs using Euclidean normalized decision matrices. Euclidean normalization prevents any parameter from growing boundlessly and thus keeps balance among the parameters for CH selection. After nominating a CH, the EPO (Emperor Penguin Optimization) algorithm has been used to detect optimal path for data delivery. EPO enhances multi-hop routing efficiency and minimizes energy utilization during data transmission, even on longer paths. The MCH-EOR is a multi-objective and swarm-based optimization technique for hierarchical WSN. It uses an unequal clustering approach to divide the network in a distributed manner. Here, the cluster size near BS is relatively small, and the intra-cluster communication is less in the clusters near BS, thus reduces overloading on nodes near BS. It eliminates the hotspot problem. The node clusters have been created using Haversine distance that reduces the dimensionality of message transmission and preserves energy. After cluster formation, CH is selected for each cluster using a probabilistic approach and a practical fitness function. The fitness function is a weighted sum approach that has been developed, keeping multiple objectives into consideration. After selecting the CH, SailFish Optimizer (SFO) is applied to find an optimal path to BS/sink node for energy-efficient transmission. The three parameters namely energy, throughput and link quality are considered to prevent sensor nodes' failure and traffic control. Instead of each sensor node's energy, the energy of CH in the path is estimated, leading to lesser energy consumption in estimation. The algorithm periodically monitors the path before starting data communication in all iterations. If any node with less than expected energy is found, it can be removed from the path, and a different path is then selected. It eliminates the problem of network holes and makes the network more robust. The LBNR-LEACH approach consider network load for candidate CHs and is estimated based on number of link, link priority and energy consumption in each round. The algorithm computes the node rank and waits for significant time to estimate the network load on available CH. After that the optimum cluster head is selected. All the proposed algorithms are compared to the existing protocols in terms of energy consumption and various Quality of Service (QoS) parameters like throughput, network lifetime, number of node alive and packet delivery ratio. The results obtained from different simulations indicate that the proposed protocols perform better than the compared protocols in terms of energy efficiency and other QoS parameters.