基于马尔科夫预测的多传感器故障检测与诊断机制

随着无线传感器网络应用规模的不断扩大,各类应用中传感器故障检测与诊断成为系统正常作业、安全可靠性保障的关键技术。针对多传感器系统与节点工作过程定义3种状态,基于故障检测信息建立状态转移矩阵,通过马尔科夫模型预测传感器故障信息,为故障检测与诊断提供决策依据。另外,拓展数据包信息字段包括故障类型、节点定位等,故障处理后节点转移至正常状态后将故障处理和诊断特征等信息存储到网关或者汇聚节点,为改善故障检测精度和诊断效率以及系统资源利用率提供依据。实验结果表明：所提故障检测与诊断算法与传统算法相比,具有更高的故障检测精度,更短的故障诊断时延、能够准确判断故障类型等性能。

Multi-sensor fault detection and diagnosis mechanism based on Markov prediction

With expanding applications scale of wireless sensor networks （WSNs） , fault detection and diagnosis become the key technologies to guarantee normal operations, security and reliability of system in a variety of sensor network applications. Therefore, define three kinds of state for multi-sensor systems, based on fault detection information establish state transition matrix. Then sensor fault information can be predicted by Markov model, which provide decision-making basis for fault detection and diagnosis. In addition ,expansion of packet information field includes fault type, node positioning, node metastasis to normal state after troubleshooting and diagnostic features and other information are stored in gateway or sink node, provide basis for improving diagnostic precision and efficiency of fault detection and system resources utilization. Experimental results show that compared with traditional algorithm, the proposed fault detection and diagnosis algorithm has higher fault detection precision, shorter fault diagnosis delay, accurate determination fault type and other properties.