Mechanical and Civil Engineering Seminar

Mechanical and Civil Engineering Seminar Series

Title: The "Underestimated" Unseen in Digital Image Correlation: Digital Twin in Material Identification, Full-Field Model Validation and Fracture Mechanics

Abstract: Recent advances in computer vision have led to the rise of highly expressive 3D scene models such as NeRFs and GSplats. More than just rendering lifelike images, these models allow robots to ground visual, semantic, physical, and affordance properties in a common 3D model, to rearrange objects in the scene and even simulate physical interactions. In this talk I will describe our efforts to build new robot autonomy features around these models, while preserving safety, modularity, and interpretability. I will present navigation algorithms for robots to safely maneuver through their environment using NeRFs and GSplats, while training the model online in a SLAM-like fashion. I will describe methods to embed semantic and affordance information into radiance fields, giving robots a 3D grounding for understanding and executing tasks from natural language commands. Finally, I will describe using these models as high-fidelity training environments for learning end-to-end visuo-motor policies. I will demonstrate such a policy for navigating a drone through an obstacle-rich environment while being robust to significant visual distractors and dynamic perturbations. I will conclude with future opportunities and challenges in 3D world models for robot autonomy.

Bio: In June 2002 Pascal Lava obtained a master degree in mathematics at Ghent University, followed by a PhD in nuclear physics at Ghent University in 2006. From January 2008 till April 2014, he worked as a professor at the MTM department at Leuven University. He is a co-founder of the spin-off company MatchID, developing optical measurement solutions or Digital Image Correlation (DIC) to assess materials and structures. Currently, he is Managing Director and CTO within MatchID with a specific focus on the integration of DIC with design and development: from the calibration of material models towards the validation of FEA simulations. His personal drive is to improve the general knowledge about DIC and its error sources, increasing the technique's range of credibility and applicability in a plethora of verticals. Pascal is the author of more than 50 peer-reviewed journal papers, a fellow and board member of the International DIC Society and received the SEM A. J. Durelli Award for outstanding contributions to DIC.

NOTE: At this time, in-person Mechanical and Civil Engineering Lectures are open to all Caltech students/staff/faculty/visitors.