针对无线传感器网络的故障特点以及故障诊断的自学习问题，本文提出一种基于免疫危险理论的无线传感器网络节点故障诊断算法。该方法利用危险触发阈值来识别危险源，用遗传算法生成抗体库，基于K近邻分类法构建多抗体故障检测器并进行故障分类，通过追踪故障数据变化更新抗体库。实验仿真证明，该算法与其他无线传感器网络故障诊断算法相比，在训练数据较少的情况下，诊断准确率更高，效率更好，耗用硬件计算资源更少，并具备动态更新特性。%Addressing to fault characteristic of wireless sensor networks( WSNs) and self-learning of fault diagnosis, an approach of node fault diagnosis algorithm based on immune danger theory for WSNs is presented in this paper. This method can identify dangerous sources with triggering thresholds of dangers. An antibody base is generated by the genetic algorithms. A multi-antibody classifier is based on the K-nearest neighbor and applied to diagnose faults for WSNs. The antibody base is updated by tracing the change of the fault data. The experiments show that the proposed method is more effective on performance of algorithms, more accurate on fault diagnosis, and less on computing resource of hardware,as well as of dynamically updating characteristics with a limit training data.
展开▼
