Sprungmarken

Servicenavigation

Hauptnavigation

Sie sind hier:

Hauptinhalt

Detailseite



Distributed optimization for cyber-physical systems

Chair: Institute ie3

Supervisior: Prof. Dr.-Ing. Timm Faulwasser, Jan Elvermann ,

Start: Sommersemester 2020 - 01.04.2020

Maximum amount of participants: 6

Description: The number of realistic industrial use cases for embedded control of cyber-physical systems using small-scale computational units is steadily increasing. Typical applications are for example Internet of Things, smart energy systems, building automation and smart homes.
In the design control solutions for such systems, numerical optimization is key to achieve coordination and consensus between the different subsystems. Applications of optimization for cyber-physical systems include distributed and decentralized Model Predictive Control (MPC), distributed state estimation and sensor placement problems

The goal of this project group is the development of a test platform for distributed numerical optimization for cyber-physical systems, which allows testing distributed numerical algorithms on embedded systems.

    Tasks:
  • Familiarize yourself with concepts for distribution optimization like the Augmented Direction of Multipliers Method (ADMM) and other methods
  • Develop a test setup that allows to couple 5-10 Raspberry Pi 3 with a central coordinating unit
  • Develop Matlab/Python tools that allow to distribute an optimization problem to the test platform
  • Prüfung der Funktionalität des Gesamtsystems
  • Validate the platform with realistic examples
  • Document and present results


    Requirements for participation (but not limited to):
  • Experience with programming (Matlab, Python, C/C++)
  • Interest and motivation for learning about numerical optimization


    Additional knowledge that is beneficial:
  • Basics of Model Predictive Control
  • Basics of numerical optimization


Depending on language competences of the participating students, the group will run in English or German.


 

The minimum amount of participants has not been reached.