Light-Weight MPC

- A lightweight simulation software implementing linear Model Predictive Control on Finite Step-Response Models

This is a repo for implementing the master's thesis for the study programme Cybernetics & Robotics at NTNU. The thesis is handed out by Equinor, and aims on implementing a simpler software framework simulating optimized control on step response models.

Project thesis description

Master student:

• Geir Ola Tvinnereim

Supervisors:

• Prof. Lars Struen Imsland (ITK)
• Dr. Gisle Otto Eikrem (Equinor)

Purpose:

The purpose of the software is to shine light on Model Predictive Controllers providing a neat and lightweight simulation tool for a JavaScript React frontend and C++ backend.

Software applications:

• Industrial use, determing the optimal MPC tuning.
• Academic framework understanding the controller method.

backend - MPC-simulator

• C/C++ module for simulating an MPC controller on an FSRM
• Python module using pyplot.matplotlib to visualise simulation data
• Emscripten compiler to compile code to Webassembly format to interface JavaScript

frontend - Web-application

• JavaScript React web application implemented using Create-React-App.

LICENCE:

Released under the terms of the BSD 3-Clause License

Master thesis

Light-Weight MPC, TTK4900 Engineering Cybernetics Master's thesis.