Achtung:

Sie haben Javascript deaktiviert!
Sie haben versucht eine Funktion zu nutzen, die nur mit Javascript möglich ist. Um sämtliche Funktionalitäten unserer Internetseite zu nutzen, aktivieren Sie bitte Javascript in Ihrem Browser.

Info-Icon Diese Seite ist nicht in Deutsch verfügbar
[Translate to Deutsch:] Foto: Judith Kraft Bildinformationen anzeigen

[Translate to Deutsch:] Foto: Judith Kraft

NICCI

Introduction

The notion of Cyber Physical System refers to Connecting Sensors, Actuators and Controllers of the control loop via a network. In this project, we strive to design a controller and network resource management unit separately, but still they exchange requirements capabilities and configurations messages. The project particularly give intention to wireless communications, where the network users are connected inside a wireless cell that is managed by radio resource management unit. One of the main challenges would be, given the limited data rate, delays and erroneous transmission, how can one guarantee safe and reliable operation of the system controlled visa wireless network?

 Objectives

  • Determine the feasibility and desirability of a close, information-exchange-driven cooperation between controllers and network resource management.
  • Develop architectural models (including interfaces, data formats, and protocol definitions) for this integration, ensuring modularity and generality rather than problem-specific design.
  • Develop core algorithms for this architecture, e.g., compression of diversity-based network capabilities meaningful for control, controllers which make best use of and adapt to changing network capabilities.
  • Integrate prediction techniques, for both control demand and channel state, in the overall system design.
  • Evaluate the resulting architecture and algorithms by analysis and simulation

Work Packages

The Computer Network Research Group of Paderborn University mainly involved.

Work Package 1

Study capacity regions for control applications; translate results to industrial control communication standards; setup the simulation environment for the network part

Work Package 2

  • Design an algorithm to compute configuration options from channel state information. Provide compact representations to use between resource management and controller. Evaluate networking aspects of overall design.
  • Modeling techniques, resource management/scheduling algorithms and tradeoff analysis (likely, simulation-based) for controller-suitable QoS specifications.

Work Package 3

  • Design and evaluation of techniques to adapt the presentation of options to the controller, based on both implicit and explicit information obtained from the controller. Detailed design and evaluation of the interaction between controller and resource manager, integration of the resulting transmission decisions into actual communication protocols. Extensive evaluation of the resulting overall system.

Publications


Liste im Research Information System öffnen

Modelling Time-Limited Capacity of a Wireless Channel as aMarkov Reward Process

B. Shiferaw Heyi, H. Karl, Proc. of IEEE Wireless Communications and Networking Conference (WCNC), 2018


DeepCAS: A Deep Reinforcement Learning Algorithm for Control-Aware Scheduling

B. Demirel, A. Ramaswamy, D. Quevedo, H. Karl, 2018

DOI


Deep reinforcement learning for wireless sensor scheduling in cyber–physical systems

A.S. Leong, A. Ramaswamy, D.E. Quevedo, H. Karl, L. Shi, Automatica (2019), 108759

In many cyber–physical systems, we encounter the problem of remote state estimation of geo- graphically distributed and remote physical processes. This paper studies the scheduling of sensor transmissions to estimate the states of multiple remote, dynamic processes. Information from the different sensors has to be transmitted to a central gateway over a wireless network for monitoring purposes, where typically fewer wireless channels are available than there are processes to be monitored. For effective estimation at the gateway, the sensors need to be scheduled appropriately, i.e., at each time instant one needs to decide which sensors have network access and which ones do not. To address this scheduling problem, we formulate an associated Markov decision process (MDP). This MDP is then solved using a Deep Q-Network, a recent deep reinforcement learning algorithm that is at once scalable and model-free. We compare our scheduling algorithm to popular scheduling algorithms such as round-robin and reduced-waiting-time, among others. Our algorithm is shown to significantly outperform these algorithms for many example scenario


Multi-agent Policy Gradient Algorithms for Cyber-physical Systems with Lossy Communication

A. Redder, A. Ramaswamy, H. Karl, in: Proceedings of the 14th International Conference on Agents and Artificial Intelligence, SCITEPRESS - Science and Technology Publications, 2022

DOI


Liste im Research Information System öffnen

Sie interessieren sich für:

NICCI

Information about the project:       
   
Project members: Holger Karl
   
Project website:

www.spp1914.de

   
Type: DFG project
Started: January 2017
Finished: Active
   
Contact: Holger Karl

Die Universität der Informationsgesellschaft