Shao Sujie, Guo Shaoyong, Qiu Xuesong
State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing 100876, China.
Sensors (Basel). 2017 Apr 28;17(5):983. doi: 10.3390/s17050983.
Due to the increasingly important role in monitoring and data collection that sensors play, accurate and timely fault detection is a key issue for wireless sensor networks (WSNs) in smart grids. This paper presents a novel distributed fault detection mechanism for WSNs based on credibility and cooperation. Firstly, a reasonable credibility model of a sensor is established to identify any suspicious status of the sensor according to its own temporal data correlation. Based on the credibility model, the suspicious sensor is then chosen to launch fault diagnosis requests. Secondly, the sending time of fault diagnosis request is discussed to avoid the transmission overhead brought about by unnecessary diagnosis requests and improve the efficiency of fault detection based on neighbor cooperation. The diagnosis reply of a neighbor sensor is analyzed according to its own status. Finally, to further improve the accuracy of fault detection, the diagnosis results of neighbors are divided into several classifications to judge the fault status of the sensors which launch the fault diagnosis requests. Simulation results show that this novel mechanism can achieve high fault detection ratio with a small number of fault diagnoses and low data congestion probability.
由于传感器在监测和数据收集方面发挥着越来越重要的作用，准确及时的故障检测是智能电网中无线传感器网络（WSN）的关键问题。本文提出了一种基于可信度和协作的新型WSN分布式故障检测机制。首先，建立合理的传感器可信度模型，根据其自身的时间数据相关性识别传感器的任何可疑状态。基于该可信度模型，然后选择可疑传感器发起故障诊断请求。其次，讨论故障诊断请求的发送时间，以避免不必要的诊断请求带来的传输开销，并提高基于邻居协作的故障检测效率。根据邻居传感器自身的状态分析其诊断回复。最后，为进一步提高故障检测的准确性，将邻居的诊断结果分为几类，以判断发起故障诊断请求的传感器的故障状态。仿真结果表明，这种新型机制能够以少量的故障诊断实现高故障检测率和低数据拥塞概率。
