Caltech Postdoc Launch Seminar

Lunch will be served at 11:45 AM (for Caltech Postdocs)

Dr. Maksym Figat

Robotic System Task-to-Model Transformation Methodology 

Designing robotic systems for safety- and mission-critical applications requires bridging the gap between high-level task descriptions and precise, analyzable specifications. This talk introduces the Robotic System Task-to-Model Transformation Methodology, an ontology-driven methodology that translates tasks into formal models of robotic systems. The approach employs stochastic timed Petri nets with resources to represent mission, system, and subsystem layers, and leverages Monte Carlo simulations to evaluate performance, robustness, and guide architectural design choices under uncertainty. A tower-building case study demonstrates how task descriptions can be systematically mapped into Petri net models that inform capability allocation, redundancy, and reliability analysis. The methodology already supports reasoning about system behavior and resource-aware design. In the future, it can also guide architectural decisions aimed at developing safe and reliable autonomous robotic systems, as well as inform the design of multi-robot missions such as NASA's CADRE, where decentralized planning and adaptive resource management are essential for planetary exploration.

For more information, please contact Sejun Kim by email at [email protected] or visit this link to fill up the form and nominate yourself as a speaker.