verify Model-based design

Overview:

Organization: QRA Corp

Product: QVtrace - Software

Role: Lead Product Designer

Industry: B2B Automotive

Toolstack: Figma, FigJam, GitHub, Adobe Suite

Introduction

Let me guide you through the design process of QVtrace, a model-based verification tool aimed at identifying and correcting errors in complex engineering systems early in the design phase.

QVtrace converts engineering models (such as those from Simulink) into formal mathematical representations, enabling rigorous checks based on data and constraints. This approach helps detect inconsistencies, flaws, and deviations from design specifications.

The Starting point

I joined as the lead product designer during the early development of QVtrace. The initial concept aimed to enhance the design process by proactively identifying flaws before they escalated into costly redesigns. The concept was admittedly unpolished, but it already defined key value outcomes.

Our business objectives:

  1. Increase Verification Efficiency: Develop a tool that simplifies system verification, accelerating design iterations and minimizing errors.

  2. Improve Constraint Management: Enable engineers to spot potential issues early by improving the review and management of design constraints.

  3. Enhance Analysis: Provide rigorous error-checking capabilities that appeal to high-stakes industries like automotive.

QVtrace's value was immediately evident through R&D collaborations, allowing us to refine the product based on real-world needs.

Original QVtrace Design Concept

understanding our users

While working with our R&D contracts, we sought to understand which types of engineers were using QVtrace and why it was critical to their work. We identified two main groups: systems engineers and safety engineers.

Systems engineers used QVtrace to test their model designs while assembling different systems. However, they struggled to access variable information within their models—they had to review the results and then manually locate the system connected to each variable.

Safety engineers, on the other hand, focused on ensuring compliance with regulations and standards by testing models using constraints. They faced challenges in managing the constraints, which were currently entered into an unorganized text box. They also recommended implementing a graphing system for analyzing time steps, as one user described: “It would be like being able to break down a car crash that never happened into each second as it occurred.”

Systems Engineer

As an automotive systems engineer, this individual is responsible for the integration and development of complex vehicle systems such as powertrain, braking, and autonomous driving features. They work on creating models that describe system behaviour, ensuring that every subsystem interacts correctly.

Why They Use QVtrace: They rely on QVtrace to verify the integration of multiple subsystems in early-stage designs. QVtrace helps them detect errors within their system models—such as inconsistencies between sensor inputs and control algorithms—before physical prototypes are built. It optimizes the design for safety, performance, and compliance with automotive standards.

Safety Engineer

This professional focuses on ensuring that the vehicle's design meets all safety regulations and standards, such as ISO 26262 for functional safety in road vehicles. They work on analyzing the design of critical safety systems like airbags, collision detection, and electronic stability control.

Why They Use QVtrace: QVtrace helps the automotive safety engineer validate that the safety-critical systems are designed correctly from the outset. By automatically checking the design for flaws that could compromise the system’s ability to meet regulatory safety constraints, QVtrace allows them to spot potential failures early, such as an error in sensor-response timing during an emergency braking event.

Linking Business Objectives to User Pain Points:

  • Objective 1: Increase Verification Efficiency

    • Users are unable to view result variables within the model to determine which subsystem contains errors.

    • Users are unable to view result variables over time to identify critical failure points.

  • Objective 2: Improve Constraint Management

    • Users are unable to view constraints simultaneously.

  • Objective 3: Enhance Analysis

    • Users can apply constraints only within a single dialogue, limiting analysis efficiency.

Ideation

During the ideation phase for QVtrace, we collaborated closely with the product and development teams to address the primary pain points of system and safety engineers. Our focus was on improving verification efficiency by optimizing how users interact with model variables. System engineers needed a faster way to identify which systems contained errors, so we implemented an overlay feature that displays variable results directly onto the model.

Another key challenge was constraint management. Safety engineers found it difficult to manage constraints in a minimized interface, where they could only apply constraints within a single dialogue. To address this, we envisioned a system that provides simultaneous visibility of constraints and allows multi-variable applications. This solution speeds up workflows, improves accuracy, and helps users comply more efficiently with regulatory safety standards.

Initial Flow Diagram

The flow diagram was centered around the QVtrace workspace, where users accessed their projects. The design was structured to present essential information upfront, with the model, variable results, and constraints visible on the dashboard. Each window addressed a specific pain point, while additional tools and options were placed in the menu to maintain a clean, streamlined interface. This approach balanced accessibility with functionality, empowering engineers to manage complex systems with greater ease.

Flow diagram of QVtrace showing project access, workspace functionality, and organizational structure.

Low-Fi Wireframes

During my low-fi prototyping, one area we focused on was how users would utilize our workspace. We aimed to quickly ideate some of the "common" layouts of windows.

This was particularly important because, for system engineers, the model held greater significance, while safety engineers were more concerned with the constraints.

We also wanted to adhere to standard workspace functionality, truly allowing users to work in a way that best suited their needs.

Example of low-fidelity wireframes for QVtrace workspace functionality, and “common” layouts of windows.

Advancing from the flow diagrams and rough wireframes, I designed the high-fidelity prototype using a colour palette that began with QVtrace's signature green. Since many users preferred a darker interface, we created a high-contrast, easy-to-read workspace where alert colours were used to highlight errors in the model's variables, drawing immediate attention to them. For safety engineers, we provided more space and organization for managing analysis constraints by placing them in a structured box system. Our goal was to avoid overwhelming the user, so we used colour strategically to emphasize critical verification information.

Design

QVtrace high-fidelity showing project access, workspace functionality, and organizational structure.

usability testing

Surveying Users

During our experiments, we aimed to ensure that users had enough space and visibility for their constraints. I surveyed our users to determine the average number of constraints applied to their models, as gathering this information within QVtrace itself would have proved nearly impossible due to the proprietary nature of models and constraints. We concluded that an average would provide a useful indication, with most users reporting between 3 to 5 constraints per analysis.

Question Posed to Our Respondents:

On average, how many constraints do you apply to your model for each analysis?

On average users analyze 3 to 5 constraints against their model.

Implementation

Development Handoff

I collaborated closely with developers to ensure accurate design implementation by overseeing key aspects of the development process. This included utilizing Figma's Developer Mode to streamline the handoff, providing developers with easy access to design specifications such as colors, spacing, and typography, and managing asset exports to optimize both clarity and performance. Additionally, I created a comprehensive design system in Figma, maintaining consistency in layout, dimensions, and spacing across different devices and screen sizes, which helped preserve the design's visual integrity across the team.

Example of assets provided via Figma's Developer Mode.

Launch Metrics

Following the launch of QVtrace, we observed key performance improvements. User engagement increased significantly, with systems engineers reporting a 25% reduction in error detection time.

Additionally, the new multi-variable constraint management feature received positive feedback from early adopters, with 90% of users stating that the interface allowed for faster and more accurate verification processes. These metrics confirm the positive impact of QVtrace on both productivity and design verification in the automotive sector.

Reflection

Problems & Solutions

  • Problem 1: Users were unable to view result variables within the model to identify which subsystem contained errors.
    Solution: Implemented an overlay feature that displays variable results directly onto the model for faster error identification.

  • Problem 2: Safety engineers could only apply constraints to one variable at a time, slowing down analysis.
    Solution: Developed a system allowing simultaneous visibility of constraints and multi-variable application to improve workflow speed and accuracy.

  • Problem 3: The disorganized interface made constraint management challenging for safety engineers.
    Solution: Designed a structured box system that organizes constraints, improving visibility and usability.

Lessons Learned

Collaborating closely with users early in the design process leads to more focused and effective solutions. By maintaining regular communication with systems and safety engineers, we gained valuable insights into their unique workflows and pain points. This approach ensured that the product was built around real-world user challenges rather than assumptions. It also allowed us to iterate more quickly and adjust features as needed based on direct feedback, fostering a user-centred design that addressed specific needs while aligning with business objectives.

Future Improvements

After the project launch, we identified areas for improvement to address in upcoming releases and on the roadmap. Offering customizable dashboard options would empower engineers to tailor their workspace to their specific needs and preferences. Engineers often prioritize different tools depending on their tasks—some may need quick access to constraints, while others focus more on model variables. Allowing users to rearrange and configure their workspace would enable QVtrace to support a wide range of workflows, enhancing user satisfaction and increasing the software’s versatility for diverse engineering roles within the automotive industry.

CONCLUSION

This case study highlights the QVtrace project, showcasing the design process and outcomes. Through research, iterative design, and testing, we successfully developed a platform that enables engineers to identify system errors and manage constraints effectively, improving their design verification and decision-making processes.