Research

We develop computational tools and smart methods to help engineers design reliable and robust products that perform as intended, even in unpredictable real-world conditions.

Engineers must account for uncertainty arising from:

Manufacturing and Materials: Variations and imperfections.
User Conditions: Unpredictable or random usage.
Machine Learning Models: Errors and uncertainty introduced when ML is used in design decisions.

Our goal is to accurately predict, manage, and mitigate the impact of these uncertainties to ensure design dependability.

Core Focus Areas

Reliability-Based Design (RBD)

Creating optimal designs that limit the probability of failure below a specific, permitted threshold.

Robust Design (RD)

Developing optimal designs whose performance is insensitive to everyday variations and fluctuations, ensuring consistent function.

Design Under Machine Learning Uncertainty

Estimating machine learning prediction errors (model uncertainty), propagating that uncertainty through the analysis, and developing numerical tools to mitigate its impact on engineering design and decision-making.