Integrated Communication and Control for Networked Systems

This direction is split into other directions: communication-control co-design. Its last version stays readable and citable here.

Version 1 of 150 papers61 conceptsfrozen 2026-10-04

In traditional engineering, control theory and network communication are treated as separate worlds. Control theorists design algorithms assuming the network is a "black box" with some fixed delay or packet loss. Meanwhile, network engineers optimize for raw throughput or latency, treating control signals as generic bits of data.

This "separation principle" breaks down in modern Cyber-Physical Systems (CPS) like autonomous drone swarms, industrial robots, and self-driving car platoons. In these environments, the network is often wireless, unpredictable, and resource-constrained. As Burak Demirel and Mikael Johansson argue in Modular design of jointly optimal controllers and forwarding policies for wireless control, siloed design leads to controllers that are too fragile for real-world glitches and networks that waste energy transmitting useless data.

The emerging research direction is Co-design: a framework where communication and control are optimized simultaneously to ensure stability and efficiency.

The Cost of Information Utility

The fundamental shift in this field is moving from "maximizing data" to "maximizing utility." In a control loop, not all packets are created equal. A robot that is perfectly on track doesn't need to send frequent updates, but a robot nearing an obstacle needs every bit of bandwidth available.

Cheng Feng and Kaibin Huang describe this as Goal-Oriented Wireless Communication. Instead of a standard schedule, the system uses "semantic communication" to transmit only what is relevant to the task. For example, Minjie Tang and Chenyuan Feng demonstrate in Decentralized Semantic Communication and Cooperative Tracking Control that a UAV swarm can maintain higher tracking accuracy with lower power by triggering transmissions based on the urgency of the tracking error rather than a fixed clock.

Channel-Aware Control

In a co-designed system, the controller isn't just a passive user of the network; it is "channel-aware." When wireless signal quality fades, a smart controller can adjust its aggressiveness.

Vinícius Lima and Alejandro Ribeiro have pioneered model-free approaches using Constrained Reinforcement Learning. By feeding real-time channel fading states directly into the control policy, the agent learns to "wait out" bad signal patches or prioritize transmissions for plants that are nearing instability. This coupling allows the system to exploit temporal variations in the environment that a decoupled system would simply see as "noise."

Similarly, Bin Hu and Philip V. Orlik show that in factory automation, coordinating physical motion with network availability—such as moving a robot arm in a way that doesn't block its own line-of-sight wireless link—can significantly expand the feasible region for safe operation.

Balancing Latency and Reliability

A core tension in networked control is the tradeoff between speed and certainty. Is it better to have a "fast but flaky" connection or a "slow but steady" one?

Konstantinos Gatsis and George J. Pappas provide a formal methodology for this in Latency-Reliability Tradeoffs for State Estimation. They argue that the "optimal" network setting depends entirely on the physical dynamics of the system. A fast-moving pendulum requires low latency even at the cost of reliability, while a slow-moving chemical process might prefer high reliability at the cost of delay.

This awareness extends to how we handle missing data. Abanoub M. Girgis and Jihong Park use Gaussian Process Regression to predict missing control signals, allowing the network to support many more devices by only transmitting when the prediction error grows too high. Even "silence" can be a signal; Shubham Aggarwal and Dipankar Maity demonstrate that if a scheduler only sends data when a value is "surprising," the absence of a message tells the estimator that the system is behaving as expected.

The 3C Convergence

The ultimate goal of this research is the "3C co-design": the seamless integration of Communication, Computing, and Control.

As Zhibo Pang outlines in the Cloud-Fog Automation paradigm, future industrial systems will not treat these as layers in a stack, but as a single unified resource. By treating network buffers as a "communication price" (as explored by Markus Klügel and Mohammad H. Mamduhi in Joint Cross-Layer Optimization), the system can dynamically reallocate bandwidth and computing power to the specific control loops that need them most at any given millisecond.

This holistic approach, moving from the decentralized frameworks of Francesco Bullo to the high-speed multi-hop guarantees of Fabian Mager and Sebastian Trimpe, ensures that as our machines become more autonomous, the "nervous system" of the network remains perfectly in sync with the "muscles" of the physical world.

Papers in this version

History & citationVersion 1 · 2026-10-04

Split into other directions · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_integrated_communication_and_control_for_94240836,
  title = {Integrated Communication and Control for Networked Systems},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/integrated-communication-and-control-for-networked-systems/rdr_942408367581402f94d3c3ad654247b6},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_integrated_communication_and_control_for_94240836_v1,
  title = {Integrated Communication and Control for Networked Systems},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/integrated-communication-and-control-for-networked-systems/rdr_942408367581402f94d3c3ad654247b6/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}