Institute for Systems Research and Department of Electrical and Computer Engineering
Advances in Information Technology have enabled the design of complex networked systems, with large number of heterogeneous components and capable of multiple complex functions. These advances have at the same time increased the capabilities of such systems and have increased their complexity to such an extent that systematic design towards predictable performance is extremely difficult if not unfeasible today. This is especially manifested in the area of cyber-physical systems, of all scales, that have become ubiquitous.
We first describe the current state of model-based systems engineering (MBSE), a system level design methodology that addresses these challenges. We then show how MBSE leads to a new methodology for designing wireless communication network protocols, which we call Component Based Networking (CBN). The new formal and model-based approach allows a systematic study of network performance and cross-layer analysis and design of routing, scheduling, MAC and PHY layer protocols. We use the method to design new and improved neighbor discovery and topology dissemination components for mobile wireless networks. We formulate the selective link-state broadcast as a graph pruning problem with restricted local neighborhood information and we develop general conditions for the distributed local policies to preserve the stable routing paths globally.
We demonstrate that the resulting algorithm, when used with the popular ETX metric, outperforms topology control mechanisms commonly used for Mobile Ad Hoc Networks such as in the OLSR protocol. The analytical tools we use include idempotent semirings, the algebraic path problem, multi-criteria tradeoff analysis. We demonstrate further, using an example from trusted routing, that logical security rules of admission control can be combined with delay performance metrics using this multi-criteria optimization framework.
We close by describing briefly fundamental research challenges and promising applications of MBSE in several areas: collaborative robotics, collaborative heterogeneous sensor networks, cybersecurity of critical infrastructures, human-machine teams and organizations, and composite trust.