Research
Advances in Physical AI are expanding what robots and manufacturing systems can do. But when a new task arrives at a manufacturing site, how do we know whether a system can perform it under the conditions that matter?
My research focuses on this gap between demonstrated performance and real manufacturing requirements. Drawing on robotic automation, digital twins, and capability representation, I study how to assess what a system can do, under what conditions, and how its capabilities match production needs.
Methods that combine digital twins, learning-based control, and physical implementation to automate robotic tasks under constrained environments and changing system conditions.
Methods that connect manufacturing requirements with supplier production records to identify suitable resources, allocate production, and respond to disruptions.
Methods to assess and compare the task-level capabilities of AI-enabled manufacturing systems, from robotics to inspection and process control, and guide their deployment.