Labs
A working notebook of physical modeling, control, and surrogate-modeling projects — mostly built in Modelica and Wolfram Language. Each card is self-contained; open the form on any project to get in touch about that specific piece of work.
Five-notebook series walking through LQR controller design from first principles, using Wolfram System Modeler for the plant and Wolfram Language for the optimization.
Why it matters Shows the math underneath the controller, not just the toolbox.
A Modelica library that connects Modelica models to ROS 2 (Robot Operating System) on Linux — publishers, subscribers, and a working hardware-in-the-loop demo.
Why it matters Bridges physics-based simulation with the robotics software ecosystem, so a controller built against a ROS topic can be developed and tested entirely in simulation before touching real hardware.
TODO — short description of what the library does and which problems it solves.
Why it matters TODO — one-line insight.
TODO — short description of the UDE work.
Why it matters TODO — one-line insight.
Continuous-Time Echo State Networks trained on Modelica simulation data to replace expensive physics models inside calibration and optimization loops.
Why it matters Matches full-model dynamics, runs orders of magnitude faster, and exports as a native System Modeler component — so the surrogate slots back into the workflow it replaced.