Wireless sensor networks are designed in such a way that transfer sensed data to base station, while a part of network is faulty. This study suggests a fault-tolerant clustering-based multipath algorithm for wireless sensor networks. We have employed a hybrid energy-efficient distributed clustering approach, to cluster nodes. Then a backup node is selected to increase the fault tolerance of cluster head node so that on completing collecting data from sensor nodes, it stores a copy of data. While collecting data in clusters, hypothesis testing and majority voting in cluster head were used to detect the fault of nodes. Finally, three paths were adopted to transfer data from source to base station based on residual energy, number of hops, propagation speed, and reliability parameters. The results of the simulation reveal that our proposed method has improved in terms of energy (6.7%), correct data (53%), data loss (4%), and delay (5.6%) compared with other algorithms.

无线传感器网络的设计方式是将感测到的数据传输到基站，而一部分网络出现故障。这项研究提出了一种基于容错聚类的无线传感器网络多径算法。我们采用了混合节能分布式集群方法来集群节点。然后，选择一个备用节点以提高群集头节点的容错能力，以便在完成从传感器节点收集数据后，它会存储数据副本。在群集中收集数据时，使用群集头中的假设检验和多数表决来检测节点的故障。最后，根据剩余能量，跳数，传播速度和可靠性参数，采用了三种路径将数据从源传输到基站。