A wireless sensor network (WSN) consists of several tiny sensor nodes to monitor, collect, and transmit the physical information from an environment through the wireless channel. The node failure is considered as one of the main issues in the WSN which creates higher packet drop, delay, and energy consumption during the communication. Although the node failure occurred mostly due to persistent energy exhaustion during transmission of data packets. In this paper, Artificial Neural Network (ANN) based Node Failure Detection (NFD) is developed with cognitive radio for detecting the location of the node failure. The ad hoc on-demand distance vector (AODV) routing protocol is used for transmitting the data from the source node to the base station. Moreover, the Mahalanobis distance is used for detecting an adjacent node to the node failure which is used to create the routing path without any node failure. The performance of the proposed ANN-NFD method is analysed in terms of throughput, delivery rate, number of nodes alive, drop rate, end to end delay, energy consumption, and overhead ratio. Furthermore, the performance of the ANN-NFD method is evaluated with the header to base station and base station to header (H2B2H) protocol. The packet delivery rate of the ANN-NFD method is 0.92 for 150 nodes that are high when compared to the H2B2H protocol. Hence, the ANN-NFD method provides data consistency during data transmission under node and battery failure.

Wireless Sensor Networks (WSNs) are becoming increasingly ubiquitous in a wide range of applications, such as agriculture monitoring, industrial automation, and healthcare. However, their operation in resource-constrained environments presents unique challenges, including data loss, node failure, and limited network lifetime, which can significantly impact their performance and reliability. This thesis investigates the challenges and existing solutions related to data loss, node failure, and network lifetime aspects of WSNs. This research proposes a hybrid Tri-Head Fuzzy C Mean Clustering Multipath Routing Protocol (THFCMRP) that helps make WSNs more reliable, consistent, and long-lasting by addressing these critical issues. The proposed THFCMRP approach is designed by a trihead fuzzy c means clustering integrated with a multi-path routing protocol approach. Initially, groups of sensor nodes are clustered using a fuzzy c means clustering approach. Then, three Cluster Heads (CHs) such as Aggregation Cluster Head (ACH), Transmitting Cluster Head (TCH), and Backup Cluster Head (BCH), are selected. These CHs include various roles and functions. The responsibility of ACH is to aggregate data from cluster members and transmit it to TCH. The information is collected by TCH and then transmitted to BS. However, BCH plays a crucial role in ensuring network consistency and data availability. The BCH helps reduce the data loss caused by TCH and Base Station (BS) failures. The BCH offers data redundancy, improves energy efficiency, and enhances network consistency and data availability by taking over for failing aggregate and transmitting cluster heads. A multi-path routing protocol is used to generate an optimal path from the TCH to the BS for data transmission. TCH uses an optimal path that consumes less energy for data transfer, thus diminishing data loss and extending the network's life. Thus, the proposed approach optimizes network performance by minimizing data loss, maintaining network performance upon any node failure incident, and extending the network's lifetime. Data Backup, Average Correct Data Packet Transferred, Data Loss, End to End Delay, Normalized Overheads, Packet Delivery Ratio, Throughput, Routing Overhead, Number of Alive Nodes, Packet Drop Ratio, and Residual Energy metrics are evaluated to assess the efficiency of the proposed approach. The packet delivery rate of the proposed algorithm is 98%, indicating that it causes lower node failure and data loss. The data loss ratio is 0.2%, meaning most packets are successfully delivered to their destination. Compared to the existing algorithms H2B2H, FTCM, HFGWO, DHSCA, and EACNF, the proposed technique achieves 83% energy efficiency. Our results demonstrated that the proposed algorithm offers better network consistency and data availability within a WSN.