Dr. Douglass' Guided Tour Through the Wonderland of Systems Engineering, UML and Rhapsody

Systems engineering is a traditionally document-intensive activity that is concerned primarily with requirements capture, high-level architecture, and the decomposition of a system into the disciplines of electronic, mechanical, chemical, and software aspects. There is growing interest in using a model-based approach, particularly with the UML, for systems engineering. There are a number of significant advantages. Since UML captures the information as models and not as artifacts, the information about the system retains its semantic content. This means that models can be verified and validated and can even be executed by simulation or animation on the host development environment or on the actual target hardware. Further, because all the information is stored in a semantic repository, all the documentation can be dynamically linked to it (we call this "dynamic associativity") so that if the information is changed in one place, those changes automatically propagate to every place the information is used. Also, because UML is a graphical language, much more information can be conveyed much more clearly than in stacks of paper documents, and views of the semantic content can be constructed to convey just the right information at the right level of detail to the stakeholders that care about that document. Next, the UML is the standard modeling language and greatly facilitates the handing off of the requirements and architecture to the various disciplines that will design the system, especially software. This hand off, and mistakes arising from the conversion from text to software models or code, is traditionally a significant source of defects as the design moves forward. Lastly, using UML is a major advantage because the system requirements and architecture are stored in the semantic repository, they can be more-or-less automatically converted into test vectors for the validation of the implemented system.

The problem in most systems engineers' minds is "How do I DO that?" when confronted with the task of moving from a textual-document based approach to a model-based approach. This paper provides a clear means for the capture of semantic information in the UML, what we mean by "models", and how to move from text-based artifacts to models and how to combine the two effectively. A step-by-step approach using a process known as ROPES/SE (Rapid Object-oriented Process for Embedded Systems/System Engineering aspects) goes through the primary activities performed by systems engineers and how they can be easily accomplished using UML and Rhapsody, including the how to plan the migration itself. These primary steps are the capturing, specifying and validation of requirements; specification of high-level architecture, definition of subsystem functionality and interfaces, mapping system requirements down to the defined subsystems, and the decomposition of the subsystems into the disciplines of electronic, mechanical, chemical, and software aspects.

View Entire Paper | Previous Page | White Papers Search

If you found this page useful, bookmark and share it on: