Resilient Sensing for Internet of Things (IoT) in Safety-Critical Applications

Since its inception, the Internet of Things (IoT) has continuously presented new opportunities that enable significant transformation of many sectors, and its widespread adoption has led to the prediction that the global total of installed and connected IoT devices by 2025 will reach over 75 billion. Today, the application of IoT is conspicuous in almost all areas of endeavor, especially in mission-critical domains that have public safety implications. These include, among others, Smart Transportation, Autonomous Vehicles, Smart Cities, Extreme Environmental Monitoring, Disaster Response, Prediction of Natural Disasters, First-Responder Support, as well as Public Health Monitoring. Consequently, it is imperative to ensure not only the accuracy of the sensed data that are fed into the IoT systems but also the resilience of the sensors that produce the data based upon which crucial decisions would be made. In this paper, we identified the inadequacies of existing research works on the resilience of the sensors widely deployed in most IoT applications. Our work then attempts to fill the gap by proposing a resilient sensing technique based on historical data of the resource-constrained sensors taken under normal, closely monitored deployment. The solution showed a promising outcome, as the sensors in our experimental system exhibited resilience against a test adversarial environment.