A Cyber-physical system (CPS) is a computer system that interacts with the physical world to perform complex functions. Many applications of CPSs like medical devices, smart grids, and robotic swarms are safety-critical systems, but their formal analysis is difficult due to their complexity. The Simplex Architecture of Sha et. al. addresses this issue by periodically monitoring the state of the system at runtime and switching the control to a verified-safe baseline controller if safety violations are imminent. The traditional Simplex approach has two significant limitations. First, its success requires creating a provably safe baseline controller.  Second, its applicability is limited to centrally controlled systems.

This talk reviews our research in addressing these limitations of the traditional Simplex approach. We propose the Black-Box Simplex Architecture, a variant of the traditional Simplex Architecture that can guarantee the safety of the overall system in the presence of an unverified---and even incorrect---baseline controller, which we treat as a black box. Next, we introduce Distributed Simplex Architecture (DSA) which widens the scope of traditional Simplex to include multiagent systems. In DSA, each agent has a local instance of traditional Simplex such that the preservation of safety in the local instances implies safety for the entire multiagent system.