NICCI: Network-Informed Control – Control-Informed network: towards multi techNology dynamICally ChangIng networks

Why It Matters

Progress in wireless communication and computation technology steadily drive the desire to interconnect control and monitoring loops via communication networks. Technological progress has enabled many new applications over communication networks, e.g., smart homes or remote monitoring and sensing devices. However, merely slow and non safety-critical tasks are currently performed via wireless communications. This is because typically network and control systems are designed independently. Therefore, in current setups the control system cannot influence the quality of service provided by the network, and vice versa. This limits the achievable performance and feasibility of the overall system consisting of the controller, network, sensors and actuators.


The question arises how to maximize the overall performance, achieve a simple modular design and guarantee safe and reliable operation given erroneous transmission, communication delays, and limited data rates. While this is possible for simple systems via custom-tailored solutions, tightly integrating and considering the communication and control requirements, it is not possible with today’s technology in applications having a large amount of control loops, network components and which might require flexible addition or removal of new components. Applications such as smart factories with a large amount of devices that need to be controlled in real-time, connected vehicles, tactile robotics – performing remote tasks with tactile sensing, as well as collaborative swarm applications, cannot be tackled by such simplistic approaches. Summarizing, the current design limitation is that the traffic demand is often considered independently from the communication capabilities and thus the control system has to simply cope with the quality of service provided by the network.


Motivated by this challenge, the overall aim of this project is:

Derive a modular solution where the network resource managers and the controllers are modularized yet they exchange information. This exchange comprises requirements, capabilities and configuration choices. The algorithms leverage predictions of both control demands and channel state, should learn from their actions, and provide performance and stability guarantees.


The project objectives are

  • Development of machine learning-based Model Predictive Control algorithms with multiple step look ahead for networked control problems with multiple communication technologies
  • Extend control algorithm to communicate system state and control information to network managers
  • Develop network scheduling algorithms for multiple communication technologies
  • Combine network scheduling and system control algorithms for safe performance

Real-world Impact

Two real-world examples are connected autonomous vehicles and smart factory automation

Future connected autonomous vehicles that require reliable and sufficiently fast information for control and decision making. These vehicles may utilise multiple communication technologies, such as WLAN, visual light communication and mmWave wireless communications. The network properties will change, as cars and platoons travel through the environment.




Consider a smart factory automation scenario involving force feedback manipulation of an object by a robot assisted with a camera system. To handle the time critical and large amount of data for the force feedback, as well as the camera information, visible light communication (VLC) together with an industrial WLAN system might be used. For the force sensitive operation the VLC might be used to achieve low latency and high bandwidth, while for simple operations WLAN communication might be sufficient. The network properties might change, as the movement of the robotic arm can influence the VLC and as the line of sight might be blocked by other objects.

The present project will provide foundational research that will enable this class of applications.



Funding / Grants

  • German Research Foundation (DFG) Grant No. 315248657 (2016 - 2023)

Chief Investigators

Other Team Members