CPS: Small: Naming, Twinning and Observing - Towards Scalable, Reliable and Resilient CPS

The Internet of Things (IoT) is increasingly used to control systems that can cause serious harm to human health or damage property, replacing more traditional hardwired systems for controlling buildings, alerting authorities to fire and gas leaks or managing rail systems. Such systems also likely operate for years, even as the physical environment changes and sensors and actuators are replaced. For example, both the Merrimack Valley gas explosion and the Notre Dame fire may have been partially caused or made worse by IoT/CPS systems that either malfunctioned or misled the operators at crucial instants. Today, programming such systems is error-prone, particularly if devices and networks can change over time. Also, testing such systems and training operators is much more difficult than for regular information technology systems, as one cannot easily reproduce dangerous or rare physical conditions such as gas over-pressure or a fire or determine what would happen if a sensor is replaced with a new model, the network is re-engineered to use different network addresses or time synchronization fails. Improved programmability and better testing will help engineers design more dependable IoT/CPS systems and predict their behavior and thus reduce the likelihood of catastrophic failures.

The project aims to improve dependability of Internet of Things (IoT) and cyberphysical systems in two ways: first, by creating a naming and directory mechanism that allows programmers to create reliable software for controlling such systems, by avoiding reliance on low-level names such as MAC (Ethernet) or network (IP) addresses and thus making it easier to reason about system properties. The project envisions creating a naming and directory mechanism that can provide a variety of logical naming mechanisms, including administrative hierarchies, geospatial relationships, civic addresses, function and access rights. The second part of the project will develop an emulator for large-scale heterogeneous IoT/CPS called FuzzIoT. This system will combine network emulation and simulation, automated fuzzing (random variation) of key network and system characteristics, and the modeling of physical components. For example, domain experts can inject the physics-driven time-dependent behavior of buildings or vehicles, either derived from models or real-world measurements. The emulator will be designed to integrate real-world components, and allow the integration of constraints and assertions. The emulator is envisioned to be useful for software development, system assurance and operator training. The project includes multiple broader impacts including the potential for integration of the research artifacts into IoT standards.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.