Parallel and Local Diagnostic Algorithm for Wireless Sensor Networks

Abstract

In wireless sensor networks (WSNs), the sensing data of nodes have spatial similarity, so the network fault diagnosis can be done by comparing the data of neighbor nodes. When sending data, the node may send erroneous data because of the interference of the signal, thereby affecting the diagnostic accuracy of the network. This paper presents a parallel and local diagnostic algorithm (PLD) for WSN. In order to avoid the problem of signal collision, this paper constructs a special diagnosis structure, which effectively avoids the influence of signal collision on node diagnosis. The algorithm can be divided into three parts: generate the candidate sub-node set, establish the fault diagnosis structure and the diagnostic test. Diagnostic test contains four rounds. The first three rounds quickly compare perceived data of adjacent nodes in parallel, greatly reduces the time of diagnosis. In the fourth round, the most reliable node is tested with the first three rounds. Simulation results show that the proposed algorithm can guarantee higher diagnostic accuracy.