CHALLENGES

Federation across different domain specific tools, and integration of data are becoming a focus for enabling collaboration and analysis, but many obstacles remain. Emerging standards like the Functional Mock-Up Interface (FMI)2 are improving simulation interoperability, while standards like the OASIS Open Services for Lifecycle Collaboration (OSLC)3 are gaining traction to improve traceability and interoperability. 

In many industries, systems engineering still relies heavily on document-centric processes. The new emphasis on digital engineering opens new opportunities, but also introduces new integration challenges. Other model-based standards such as the Systems Modeling Language (SysML) and the Unified Architecture Framework (UAF) are continuing to evolve to provide a standard way to support model-based systems engineering for systems and enterprises.

2. https://fmi-standard.org/

3. https://www.oasis-open.org/committees/oslc-domains/charter.php

Software Complexity, Agility, and Scale

Over the last 50 years, software has become a more important component of many systems. As software has grown in scale, complexity, and interconnectivity, the software engineering community has adapted—and developed new approaches with an emphasis on agility, evolution, development and operations (DevOps), and continuous development, integration, and deployment. 

The systems engineering community is working with the software engineering community to generalize these approaches to cyber-physical systems (CPS), to bring new value and capabilities to users sooner and more often, and to balance risks, regulatory issues, and societal impacts.

Impact of AI and Autonomous Systems

Traditional systems engineering tools and practices do not address complex systems that continue to learn and modify themselves during operation. In addition, these systems can have significant social and ethical implications that need to be considered as part of the design. Examples include critical decisions that are allocated to autonomous vehicles that may impact safety, and how these systems can potentially create information that violates privacy. Methods to curate data, to assure it is not biased, and to ensure that the training sets adequately span the operational environment are only being developed now. Furthermore, validation and verification of these systems is currently based on traditional systems engineering approaches, but new techniques and concepts may be necessary to account for the opacity of how these systems make decisions. The need for ongoing verification of continuously learning and evolving systems also needs to be addressed.