Communication and control co-design for beyond 5G networks
Thesis event information
Date and time of the thesis defence
Place of the thesis defence
L10, University of Oulu, Linnanmaa
Topic of the dissertation
Communication and control co-design for beyond 5G networks
Doctoral candidate
Master of Science (M.Sc.) Abanoub Mamdouh Girgis Pipaoy
Faculty and unit
University of Oulu Graduate School, Faculty of Information Technology and Electrical Engineering, CWC-Radio Technologies
Subject of study
Communications Engineering
Opponent
Professor Gilberto Berardinelli, Aalborg University
Custos
Professor Mehdi Bennis, University of Oulu
The thesis proposes a novel communication-efficient control framework to enable robust and scalable wireless control under limited network resources.
The advancement of wireless communication has revolutionized the design of flexible and spatially distributed control systems, offering significant advantages over traditional wired architectures. However, integrating wireless networks into control loops introduces communication delays and unreliability, which can degrade control performance. While ultra-reliable low-latency communication (URLLC) meets stringent control requirements, it is resource-intensive and limits scalability. Conversely, massive machine-type communication (mMTC) supports scalability but often degrades control performance. These limitations stem from the fact that technological developments in communication and control are carried out in silos, highlighting the need for a joint communication and control design.
The thesis proposes a novel communication-efficient control framework to enable robust and scalable wireless control under limited network resources. First, an age-of-information (AoI)-aware scheduling and power allocation scheme with two-way Gaussian Process Regression (GPR) is introduced to update the most critical control system while predicting future states and control commands, improving control stability and scalability. Next, a two-way split Koopman auto-encoder framework is proposed for remotely controlling a non-linear system under limited wireless resources by predicting state and command information at the controller and actuator, respectively. To further reduce computation complexity and communication overhead in non-linear systems, we propose a semantic Koopman-based communication and control co-design framework. This includes the proposed compositional logical dynamical (CLD)-Koopman auto-encoder for facilitating compositional control across correlated control systems. Finally, a time-series joint embedding predictive architecture (TS-JEPA) is proposed to control vision-based systems under bandwidth constraints by encoding high-dimensional frames into semantic embeddings and predicting their future evolution, supported by a channel-aware scheduler that prioritizes critical transmissions based on AoI and channel conditions. The simulation results validate the effectiveness of the proposed approaches in achieving real-time, robust, and scalable control across large-scale wireless systems.
The thesis proposes a novel communication-efficient control framework to enable robust and scalable wireless control under limited network resources. First, an age-of-information (AoI)-aware scheduling and power allocation scheme with two-way Gaussian Process Regression (GPR) is introduced to update the most critical control system while predicting future states and control commands, improving control stability and scalability. Next, a two-way split Koopman auto-encoder framework is proposed for remotely controlling a non-linear system under limited wireless resources by predicting state and command information at the controller and actuator, respectively. To further reduce computation complexity and communication overhead in non-linear systems, we propose a semantic Koopman-based communication and control co-design framework. This includes the proposed compositional logical dynamical (CLD)-Koopman auto-encoder for facilitating compositional control across correlated control systems. Finally, a time-series joint embedding predictive architecture (TS-JEPA) is proposed to control vision-based systems under bandwidth constraints by encoding high-dimensional frames into semantic embeddings and predicting their future evolution, supported by a channel-aware scheduler that prioritizes critical transmissions based on AoI and channel conditions. The simulation results validate the effectiveness of the proposed approaches in achieving real-time, robust, and scalable control across large-scale wireless systems.
Last updated: 13.10.2025