This research focuses on developing concepts and methodologies to evaluate and design property-preserving interfaces for safety-critical systems. These are systems that require real-time sensing/control and rely on existing networking infrastructures for coordination and information exchange. These systems, which range from power distribution networks and air (or other) traffic control systems to avionic controllers and to embedded automotive electronics, are characterized by high complexity that is associated with both discrete and continuous aspects. The resulting intricate behavior of these safety-critical interconnected systems challenges traditional notions for safety, security, and reliability. It necessitates the development of new methodologies for understanding how to obtain compositions of such modular systems and how to design interfaces that achieve not only robust operation and performance but also ensure trust and privacy among users. The research seeks a unifying and multifaceted approach to this problem. The approach is to decompose the large body of research on modular systems and interface design to address: state estimation in interacting modular systems and implications to safety; model verification and property checking for modular hybrid systems; and compositional models and interface design in switched discrete and continuous control systems. The research draws on areas as diverse as distributed algorithms, robust and fault-tolerant design, hybrid system control, performance evaluation, applied probability, graph theory, distributed estimation, and formal methods. These are applied to the problem of analyzing and understanding the tradeoffs involved in the design of modular systems with real-time sensing and control capabilities. The research is expected to have significant impact in permitting and enabling the ubiquitous use of critical network infrastructures for a variety of diverse applications.
