Reliability Estimation of Complex Systems Based on the Internet of Things

Large-scale medical equipment is critical in monitoring patient health conditions and making diagnoses. But when anomalies in medical equipment emerge, medical attention for patients can be delayed unanticipatedly and equipment require expensive maintenance services, which can even lead to medical accidents. This paper develops a data-driven predictive maintenance model for Computed Tomography (CT) equipment based on the Internet of Medical Things (IoMT) in the West China Hospital of Sichuan University. First, for data pre-processing step, the time interval for time series data is unified and the missing data is interpolated. Second, the sliding time window is applied in the feature construction step to combine the nearby data and form a new time series that contains historical information. Third, Discrete Fourier Transformation (DFT) and Discrete Wavelet Transformation (DWT) are applied in the feature transformation step to obtain the features in frequency and time-frequency domain. The time domain features are also extracted by computing the mean and standard deviation values. Finally, the multiple train-test splits method is applied to the selected features. The model for predictive maintenance proposed in this paper can help hospital equipment maintenance teams reduce the risk of anomalies in CT equipment.