Ubi-Asthma: Design and Implementation of Asthmatic Patient Monitoring System in Ubiquitous Geospatial Information System

In recent years, the growth of population, increase in environmental pollution, and changes in human’s life style, a dramatic increase in the number of asthmatic patients has been observed. In addition, the Lack of suitable solution for the cure of asthma shows us the necessity of recent researches for controlling of the exacerbation of asthma in order to improve patients’ quality of life. Environmental condition in a spatio-temporal perspective enables us to do some predictive analytics in order to providing this ability to the asthmatic patients to have self-checking in addition to the medication by personal physician. Ubiquitous patient monitoring system which is enabled by geospatial perspective can assist us to provide such services to the asthmatic patients in both outdoor and indoor environments. In this paper, we monitor ubiquitously asthmatic patients in a Ubiquitous GIS environment by considering 8 important environmental asthma triggers including CO, NO, O3, PM10, SO2, Temperature, Humidity, and Pressure, and 4 medical records of current status of asthmatic patients. In fact, to achieve a better modeling with divided our research to two parts: User Model and Contextual Model. For implementation of Ubiquitous Asthma model (UbiAsthma), this survey is done by consideration of 30 patients in Tehran city. To develop our prediction model in UbiAsthma, at first we used VIKOR method to reclassified current patients’ medical classification to a novel and applicable classification for providing context-aware services. Secondly, after classification of patients, one patient is selected and the information related to FVC and FEV1 of the patient in 86 different locations within 47 days is collected. The patient was equipped by CO pollutant sensor (MQ9) and Temperature, Humidity, and Pressure sensors collection by smart-phone in order to having of a real-time database for the patient’s trajectory. Also, we utilized O3, NO, SO2, and PM10 for all 86 locations as static database in our model. After data collection and manipulation, Artificial Neural Network (ANN) is used for doing predictive analysis for asthmatic patients’ status. The result of test and train of ANN method shows that UbiAsthma model has 0.0230 evaluation error in prediction of asthmatic patients’ status, which is a considerable output in our model.