嵌入式网络传感器多路观测数据实时检测仿真

Simulation of Real-time Detection for Multi-Channel Observation Data in Embedded Network Sensor

对嵌入式网络传感器多路观测数据进行检测能够实现网络传感器运行状态的及时监测,对保证网络传感器的正常运行具有重要意义。针对当前网络传感器观测数据检测方法存在的检测准确度不高,实时性较差,且影响传感器运行速度的问题,提出一种基于模糊ARHMM的嵌入式网络传感器多路观测数据实时检测方法,通过对嵌入式网络传感器观测数据进行关联规则挖掘,并分析关联规则集之间的相似度,对多路观测数据进行挖掘。利用模糊ARHMM,通过计算三部分检测的隶属函数,将挖掘的数据进行判别,通过计算网络传感器多路观测数据中正常数据集的隶属度函数,并利用三部分检测的隶属函数,实现嵌入式网络传感器多路观测数据实时检测。实验结果表明,利用所提方法进行观测数据检测,检测的准确度较高,且检测的实时性较好,对传感器运行质量影响较小。

A real-time detection method for muhichannel observation data in embedded network sensor based on fuzzy ARHMM was proposed.By mining association rule of observation data in embedded network sensor and analyzing the similarity degree between association rule sets,we mined muhichannel observation data.Then,we used fuzzy ARHMM to calculate the membership function of three-part detection and discriminated the mined data.In addition,we calculated the membership functions of normal data set in the multichannel observation data of network sensor.Finally,we used the membership function detected by three parts to realize the real-time detection for multichannel observation data in embedded network sensor.Simulation results show that the proposed method has high accuracy of detection and good real-time performance to detect observation data,which has less impact on the run- ning quality of sensor.