Publication

Abstract : In wireless sensor networks (WSNs), failures are inevitable due to the dynamic environment and ubiquitous deployment. Node failure can lead to loss of network connectivity, which can lead to data loss between two nodes. When one or more nodes fail, the failed nodes cannot send data to the currently assigned sensor nodes. Additionally, node failures affect the efficiency of data transfer between nodes based on service quality (QoS) measurements. This requires a better fault tolerance system, which can maintain quality of data transfer in the event of faults. In this paper, we propose a multipath routing algorithm using hybrid optimal fault tolerant system for WSN (HOFT-MP). First, we introduce a modified teaching–learning-based optimization (MTLO) algorithm for efficient clustering which groups the sensor nodes to improve energy efficiency. Here, we combine teacher learning with the fish swarm optimization (FSO) to increase the searching range in network which effectively computes node location, position and movement direction of sensor nodes. Second, we develop a nonlinear regression based pigeon optimization (NR-PO) algorithm to compute the backup node for clusters to detect node faults which increase the fault tolerance. Then, a deep Kronecker neural network (DKNN) is used to compute optimal path among multipath which enhances quality of data transfer. Finally, the performance of proposed HOFT-MP routing algorithm compared with the existing state-of-art routing algorithms in terms of energy consumption, end-to-end delay, throughput, data loss and network lifetime.