Below you will find descriptions of some of the open-source code developed throughout my PhD research.

UAV Digital Twin: Probabilistic Graphical Model

An open-source python and ROS2 implementation of our probabilistic graphical model framework for Digital Twins. Specifically, this repository contains a ROS2 package that implements a simulated asset and its associated digital twin.

The digital twin leverages a probabilistic graphical model to perform sequential Bayesian inference for data assimilation, which enables the digital twin to update its internal computational models of the associated asset in response to incoming sensor data. The digital twin also uses the graphical model to perform planning via model-based reinforcement learning. This process results in a control policy that enables the digital twin to select optimal control inputs based on its up-to-date computational models of the asset.

This particular implementation is tailored toward the UAV structural health monitoring use-case presented in our paper, but can serve as a starting point for other applications.

UAV Digital Twin: Experimental Data and Bayesian Model Calibration

This repository contains the experimental datasets and code used for data processing and Bayesian calibration of our UAV digital twin model. This calibration procedure is formulated in terms of our probabilistic graphical model framework for digital twins.

An open-source python wrapper for ASWING. ASWING is a program developed by Professor Mark Drela (MIT) for the aerodynamic, structural, and control-response analysis of aircraft with flexible wings and fuselages of high to moderate aspect ratio.