A Framework for Automatic Dynamic Constraint Verification in Cyber Physical System Modeling Languages
AuthorBunting, Matt Robert
Model Based Engineering
AdvisorSprinkle, Jonathan M.
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractDesign of Cyber-Physical Systems (CPSs) involves overlapping the domains of control theory, network communication, and computational algorithms. Involving multiple domains within the same design greatly increases the system complexity. Furthermore, the physical nature of CPSs generally involves important safety constraints where constraint violations can be catastrophic. The design of CPSs benefits from focusing on the construction of abstracted, high-level models in a DomainSpecific Modeling Language (DSML). A Domain-Specific Modeling Environment (DSME) may aid in the design of such complex systems by enforcing structural design constraints during the construction of models. Models built using a DSME may also use compilers or interpreters to produce real working, low-level artifacts that represent the high-level design. Though each model in a DSME may abide by a formal specification, the behavior of a design may violate dynamic constraints if deployed. Engineers are tasked to ensure that models behave safely by implementing their expert knowledge after using appropriate verification tools. Constraint violations may be eliminated by a modification of the model based on verification feedback, known as Dynamic Constraint Feedback (DCF). Mending such constraint violations is a task generally performed by the model designer. Such a process could potentially be automated through the capture of well-known design practices. The challenging task when automating model correction then becomes in the design of a DSML. A designer of a DSML may have a clear understanding of how to design the syntax and semantics for their domain, but there are no formal methods for implementing verification tools for automatic model correction. Such a framework could greatly aid in the selection of available verification tools, implement well-established design methods, and model dynamic constraints. Presented is the Dynamic Constraint Feedback Metamodeling Language (DCFML), a new metamodel to implement DCF upfront in DSML design. This particular solution provides a concrete solution to the abstraction of the various components of DCF, and then appends them to the DSML design process provided by a DSME.
Degree ProgramGraduate College
Electrical & Computer Engineering