A distributed fault recovery method in clustering-based Wireless Sensor networks

Wireless Sensor Network (WSN) is fault-prone given the resource-constrained characteristic of sensors and harsh deployment environment. Failure of sensors may cause negative consequences on the considered applications. Therefore, WSN should be able to tolerate the failures of sensors and self-recover from fault to guarantee the service. However current studies mostly focus on fault diagnosis, and little research has been done on fault recovery. In this paper, we propose a Distributed Fault Recovery Approach (DFRA) for the networks´ self-restore in case of failures with minimized calculate and communication overhead. Fuzzy theory is introduced to model the fault-recovery problem. To find an optimal replacement scheme, genetic algorithm and simplex method are adopted and converged. A fusion algorithm named genetic simplex method algorithm (GSMA) is proposed. At last, the simulation shows that GSMA has better performance and achieves longer network lifetime through the execution of the fault recovery approach.