Software

I have extensive professional experience in software engineering in Matlab/Simulink, C/C++, and Python for control-systems, data analysis, and simulation. Below are a couple of Matlab toolboxes I’ve developed:

stepalize

A robust method for building linear, time-invariant models from step responses that can be used to accurately identify time constants of high-order linear systems.

• Identifies arbitrarily high-order models without numerical differentiation or deconvolution.
• Allows for model-order selection at run-time.
• Supports overshoot/undershoot constraints via quadratic programs.
• Supports eigenvalue constraints via semidefinite programs.

The meaty details and an interesting application can be found in the forthcoming publication “Thermal Dynamical Identification of Light-Emitting Diodes by Step-Based Realization and Convex Optimization,” to appear in IEEE Transactions on Components, Packaging, and Manufacturing Technology sometime in the near future.

RBIS

Some realization-based methods for building linear, time-invariant models form arbitrary input-output data using some results from classical realization theory. This is an implementation of most of the algorithms found in my thesis.