Fault Detection And Data Reconstruction Of Body Sensor Networks Using Bayesian Network Model

The body sensor network(BSN)is limited in resources,vulnerable to environmental effects and malicious attacks,which can cause faulty sensor data,and even lead to false alarms and false medical diagnosis.Therefore,the faulty sensor data should be detected and reconstructed as much as possible before the data collected by the sensor is used for medical diagnosis.After that,more accurate judgments can be made based on the reconstructed sensor data.Using a small number of sensors to collect data is one of the advantages of the Bayesian network.First of all,a Bayesian network model is established by the temporal and spatial relationship of the collected data in the thesis.In order to improve the correctness of the collected data,the concept of the truth value of data is introduced in the model,which is added to make the sensor data credible.As the correctness of the data collected by the sensor can not be guaranteed,the truth value and the weight are added for each sample to ensure the authenticity of the data.Then the mathematical expression of the performance indicators such as true positive rate(TPR),false positive rate(FPR),inaccuracy(IACC)and false reconstruction rate for detected faults(FPPF)are obtained after a series of derivation by using the properties of Bayesian network.After that,the optimal threshold of the scheme is calculated according to IACC.Finally,the specific schemes of fault detection and data reconstruction are given.Assuming that the truth value is the same as the value of the collected data,the relationship between the conditional probability formula calculated from the fault detection algorithm and the optimal threshold can be used to determine whether the data collected by the sensor is correct or not.After analysing the simulation results of fault detection and data reconstruction scheme based on Bayesian network and other algorithms such as KNN,MD and J48&LR,the experimental results show that the correctness of the proposed scheme is higher.