Research On WSN Node Fault Detection Method Based On Belief Rule Base

Wireless sensor network(Wireless Sensor Network,WSN)is widely used in industrial monitoring,urban transportation and smart home with its advantages of high accuracy,low latency,and wireless communication.Because the failure of the WSN node causes the monitoring data to fluctuate or the monitoring fails,which affects the status judgment of the monitored object,it is necessary to perform accurate and timely fault detection on the WSN node.After studying the field of WSN sensor node fault detection,it is found that the accuracy of fault detection will be affected by uncertain factors,including the uncertainty of sensor data and the ambiguity of expert knowledge used in fault detection,which in turn affects the results of fault detection.Traditional methods for fault detection of WSN nodes mainly use a large amount of data training or the construction of expert knowledge models,which cannot take into account the uncertainty of data and the ambiguity of expert knowledge,thus reducing the generalization and practicability of the fault detection method.Therefore,in the research of WSN node fault detection methods,it is necessary to consider the influence of data uncertainty and expert knowledge ambiguity on node fault detection.In order to solve the problem of uncertain factors,this article has carried out research in the following two aspects.1)When a sensor node fails,different types of failures have different effects on node data.Model training using only raw node data has problems such as large data volume and incomplete coverage of fault types.Therefore,it is necessary to check the original node data.On the basis of feature extraction.This paper analyzes the working principle of the wireless sensor network and the working characteristics of the sensor nodes,and extracts the characteristics of the sensor data from the two dimensions of space and time.Determine the two data attributes of time correlation and space correlation to describe the numerical relationship between sensor nodes,and give the calculation methods of time correlation and space correlation.2)In order to solve the problem of inaccurate fault detection caused by uncertain factors,this paper applies the Belief Rule Base(BRB)model to the WSN node fault detection problem,and builds the BRB based on the data correlation between WSN nodes Fault detection model.Then a constraint covariance matrix adaptation evolution strategy algorithm based on projected operation(P-CMA-ES)is used to define the parameter optimization method of the model.Construct experiments through Intel Lab Data dataset to verify the effectiveness of the model,and establish comparative experiments to verify the superiority of the model.In order to verify and popularize the model in this paper,this paper develops a prototype of the WSN node fault detection system,which includes two parts: embedded data processing module and host computer application software.Use Visual Studio 2017 to build the upper computer application software,and use the STM32 chip to realize the embedded data processing module.In order to realize the rapid response of the system,the confidence rule library code is implemented in C language and solidified into STM32.The embedded data processing module is used for node status detection.The detection results and related indicators are sent to the upper computer application software through wireless communication or wired communication.The upper computer application software has the functions of human-computer interaction and real-time display of the detection results.The user can pass the upper computer application software sets the model parameters.The experiment in this paper proves that the BRB model has a good performance in the fault detection of WSN nodes,which effectively reduces the influence of uncertain factors on the fault detection of WSN nodes,and improves the accuracy of detection.The prototype of the WSN node fault detection system developed by BRB in this paper confirms the practical significance of BRB in engineering practice.