COVID-19 - An intelligent system for contact tracing, monitoring, and privacy preserving data analytics during the COVID-19 pandemic

COVID-19的迅速传播导致采取严厉措施隔离疑似感染的个人，以确保社区安全。目前，建议疑似患者在家中自我隔离，这可能无法跟踪和确认合规性。尽管一些国家通过智能手机应用程序确保执法，但只要把智能手机放在家里，就可以规避这些措施。瑞尔森大学和多伦多大学的研究人员与Dapasoft公司合作，将开发一个混合系统（一个轻便的不可拆卸的腕带，附带移动的应用程序），用于自愿或经授权的指定跟踪被建议或受自我隔离期限制的人员。智能手机的地理定位功能，加上与腕带的一致连接，将创造一个难以打破的监控系统。该系统还可以选择收集生物医学数据，如心率、体温和自我报告的数据，如呼吸道症状、与他人接触的时间和持续时间。将使用新型人工智能算法分析收集的数据（获得参与机构的适当伦理批准），以衡量主要和次要结局。该项目的主要成果是衡量遵守情况，将通过确定个人的行动模式来计算遵守情况。次要成果包括疫情风险评估和疫情集群的可视化地图。我们将采用新的信息融合机制，结合我们最近提出的无监督聚类方法来生成直观的聚类图可视化，这将非常有利于非技术领域专家（例如医疗从业者）识别数据中的趋势。我们计划将这些算法打包到一个易于部署的分析和可视化工具中，以便在未来爆发类似疫情时使用。

The rapid spread of COVID-19 has resulted in taking drastic measures in isolating individuals suspected of infection to ensure community safety. Currently, suspected patients are advised to self-quarantine at home, which can be impossible to track and confirm compliance. Although some countries have ensured enforcement through smartphone apps, such measures can be circumvented by simply leaving the smartphones at home. Researchers from Ryerson University and the University of Toronto, in partnership with Dapasoft Inc., will develop a hybrid system (a lightweight non-detachable wristband with accompanying mobile app) for voluntary or authorized assigned tracking of persons who are advised or subject to a self-quarantine period. The geolocation feature of smartphones, combined with consistent connection with the wristband, will create a monitoring system that is difficult to break. The system will also have options to collect biomedical data such as heart rate, temperature, and self-reported data such as respiratory symptoms, time and duration of contact with others. The collected data (with appropriate ethics approval from participating institutions) will be analyzed using novel artificial intelligence algorithms to measure primary and secondary outcomes. The primary outcome of this project is measuring compliance, which will be calculated through identifying the movement patterns of individuals. Secondary outcomes include risk assessment of outbreaks, and visualization maps of clusters of outbreaks. We will employ novel information fusion mechanisms, combined with unsupervised clustering approaches that we have recently proposed to generate intuitive visualization of clustering maps, that will be highly beneficial for non-technical domain experts (e.g. healthcare practitioners) to identify trends within the data. We plan to package the algorithms in an easily deployable analytics and visualization tool that can potentially be utilized in case of a similar future outbreak.