Building a complex product is challenging; understanding the full system can be even harder.

As products become more software-driven, engineering teams must manage growing numbers of requirements, interfaces, safety considerations, and design decisions. A small change in one area can affect work happening elsewhere in the system. Keeping everything aligned often requires more effort than teams expect. That challenge is what led Dr. Johannes Reinschke to start Automated Conformance.

Automated Conformance was founded to help engineering teams better understand how changes affect complex systems while reducing the manual effort required throughout development.

When Complexity Outgrows the Process

Reinschke has seen a recurring theme throughout more than 20 years in product development and engineering. Architecture development, requirements management, safety analyses, and testing often happen in separate environments. Teams may spend significant time moving information between tools, recreating existing data, and keeping documentation up to date as designs evolve. “The information is essentially the same information, just presented differently,” explains Reinschke. “Having to copy it manually into another representation is a waste of time and very error-prone.”

As projects grow larger, the potential consequences become more significant. A design change made in one area can affect multiple aspects elsewhere in the system. As Reinschke describes, an engineer can make a change but can’t be certain that this change is not introducing bugs elsewhere.

Without a clear understanding of those dependencies, teams risk introducing inconsistencies that may not be discovered until much later in development. “There’s no point in launching your product quickly if it doesn’t work as intended,” says Reinschke. “You want to make sure that product-related risks are minimized. The time savings come on top.” For companies building safety-critical products, that risk can be especially costly.

Using Models as a Starting Point

Automated Conformance approaches this challenge through model-based development. The autoConform software is designed to help engineering teams connect activities that are often spread across separate tools and processes. Rather than requiring engineers to recreate the same information across multiple environments, the software uses a system architecture model as a foundation for development activities throughout a project.

For users, the process begins with a Simulink model that captures a product’s architecture, including hardware components, software components, interfaces, and functional relationships. The team uses that model as a foundation for activities such as requirements definition, design specifications, safety analyses, and test specifications. The software can generate engineering work products from that model while helping teams maintain traceability of how specifications connect across a project.

The autoConform software also helps users manage change throughout development. The process is designed to be guiding, helping engineers understand what should happen next and which activities may need to be revisited following a change. If a modification affects downstream activities, the system identifies those dependencies and guides engineers through areas that may require review.

According to Johannes Reinschke, the goal is not simply automation. It is helping engineers understand the impact of change across increasingly complex systems.

Building on Existing Engineering Workflows

The autoConform software is developed in MATLAB and built around Simulink models. For Reinschke, who has used MATLAB and Simulink since completing his doctoral research in control engineering, the tools provide a natural foundation for the platform.

The software relies on Simulink models as a source of architectural information. Using MATLAB and Simulink APIs, the autoConform software extracts information from those models, generates engineering work products, and checks that information remains aligned as projects evolve. The ability to move information between models and generate documentation is an important part of maintaining consistency throughout product development.

The decision to integrate with Simulink was also practical. Many organizations developing safety-critical products already rely on model-based development. Rather than introducing another standalone tool, Automated Conformance builds on workflows engineers are already using. “People are already doing model-based software development with code generation and deployment based on Simulink models,” explains Reinschke. “Rather than introducing another tool at the system level, doing it the same way at the system and software level is very helpful.”

Launching Ahead

After several years of development and customer feedback, Automated Conformance is preparing for its next major milestone: product launch.

Like many technical founders, Reinschke’s role has involved more than building the product itself. It has also meant translating engineering ideas into something customers can use, evaluating feedback, and refining the experience along the way. “You have to be an optimist and you have to have a long breath,” shares Reinschke. “Obstacles appear that you had not thought about, and it may take longer than you hoped. But you carry on.”

Automated Conformance joined the MathWorks Startup Program during its development. For the small team, the biggest benefit was access. “As a startup, we are not rich. We need to be careful about our spending,” Reinschke says. “Having reduced costs in the MATLAB licenses is very helpful.” The program allows the team access to a broader set of tools while refining the product and exploring different approaches.

For Reinschke, the most rewarding part of the journey has been seeing ideas translate into practical improvements for engineers. “Seeing how things can change, where what you previously had to do in a rather awkward manner all of a sudden becomes much smoother, that’s really the fun part,” Reinschke concludes.

See the autoConform platform in action: