| Abstract: | Wireless Sensor Networks (WSNs) have energy-constraints that restricts to achieve prolonged network lifetime. To optimize energy consumption of sensor nodes, clustering is one of the efficient techniques for minimization of energy conservation in WSNs. This technique sends the collected data towards the SINK based on cluster head (CH) nodes that leads to the saving of energy. WSNs have been faced a crucial issue of fault tolerance and the overall data communication is collapsed due to the failure of cluster head. Various fault-tolerance clustering methods are available for WSNs, but they are not selected the backup nodes properly. The backup nodes’ closeness or location to the other remaining nodes is not considered in these methods. They may increase network overhead with the backup nodes accessibility. A fault-tolerance cluster-based routing method is presented in this paper that aims on providing fault tolerance for relay selection in addition to the data aggregation method for clustered WSNs. The proposed method utilizes backup mechanism & the Particle Swarm Optimization (PSO) to achieve this. Based on the distance from sink, residual energy, and link delay parameters, the CHs are chosen and the network is categorized into the clusters. The Backup CHs are selected by estimating the centrality among the nodes. As a part of intra-cluster communication for reducing the aggregation overhead among CHs, the Aggregator (AG) nodes are deployed in every cluster. So that they act as the bridge between the member nodes and CHs. These AG nodes aggregates the information from member nodes and deliver it to the CHs. The PSO with modified fitness function is used to identify the best relays between AG and member nodes. The proposed mechanism is compared with existing techniques such as EM-LEACH AI-Sodairi and Ouni (2018), QEBSR Rathee et al. (2019), QOS-IHC Singh and Singh (2019), and ML-SEEP Robinson et al. (2019). The simulation results proved that the proposed mechanism reduces overhead by 55% and improve the energy consumption & throughput by 40% & 60% respectively. |
