论文部分内容阅读
温室无线传感器网络中故障节点会产生并传输错误数据,不仅消耗节点的能量和带宽,而且导致错误决策。针对此问题研究一种准确判断节点故障状态的方法。采用时序分析和遗传BP神经网络,建立基于时间序列和神经网络的传感器节点故障诊断系统,通过对传感器样本数据进行时序分析,提取模型参数作为特征向量,并以此对遗传BP神经网络进行网络训练,实现传感器节点故障的诊断。试验结果表明:该方法能够有效地识别传感器节点故障类型,15组测试样本的输出矢量与同类故障基准矢量的欧式距离和为0.007,识别正确率为100%。
Fault nodes in greenhouse wireless sensor networks generate and transmit erroneous data, which not only consumes the energy and bandwidth of nodes, but also leads to wrong decisions. Aiming at this problem, a method of accurately judging node fault status is studied. Using time series analysis and genetic BP neural network, the sensor node fault diagnosis system based on time series and neural network is established. By analyzing the sensor sample data in time series, the model parameters are extracted as eigenvectors, and then the genetic BP neural network is trained by network , To achieve the sensor node fault diagnosis. Experimental results show that this method can effectively identify the sensor node fault types. The Euclidean distance between the output vector of the 15 test samples and the benchmark fault vector of the same type is 0.007, and the recognition accuracy is 100%.