Multiple Contributions of ComNets Dortmund at IEEE INFOCOM 2026

The Chair of Communication Networks (ComNets) at TU Dortmund University presented four contributions at the IEEE Conference on Computer Communications (INFOCOM) 2026 in Tokyo, Japan, one of the most prestigious conferences in the field of computer networks and communications. With a workshop paper, two demo contributions, and a poster, the ComNets team presented a broad spectrum of current research results spanning network co-design of Open 6G Radio Access Networks, AI-driven network planning, energy efficient IoT communications, and ad-hoc localization in applications like industrial robotics and emergency response.

O-RACES: Proactive AI-driven Scheduling in Open RAN for 6G-Networked Humanoid Robots
Real-time control of robots over wireless networks poses some of the most demanding communication requirements on future 6G systems. Here, bipedal humanoid robot locomotion relies on millisecond-level feedback loops between sensing,  centralized AI-based controller running on an edge cloud, and action, leaving only little room for latency and jitter. O-RACES (Open RAN Real-time AI-Coordinated Efficient Scheduling) addresses this challenge with a proactive AI-driven scheduling framework built on open interfaces and designed for seamless integration into the Open RAN architecture as a dApp. By predicting uplink resource demands in real-time, O-RACES pre-allocates radio resources, eliminating queuing delays while avoiding the spectrum waste of static over-provisioning. Evaluated with up to 16 simultaneously controlled virtual humanoid robots over a physical Open RAN testbed under full network load, O-RACES reduces uplink latency and spectrum consumption compared to reactive and static scheduling approaches, while maintaining stable robot locomotion even under challenging terrain and interference conditions. 

The full measurement dataset is publicly available at: 
https://doi.org/10.17877/TUDODATA-2026-MLFQ463Q

Hybrid Active-Passive Network Design Enabled by Physics-Informed ML-based Propagation Modeling 
Deploying millimeter-wave networks in industrial indoor environments remains one of the central challenges due to their complex propagation behavior, making careful network planning indispensable. A first-of-its-kind end-to-end software framework addresses this challenge by combining a Physics-Informed Neural Network for accurate path loss prediction with an intelligent placement routine that jointly optimizes the deployment of active base stations and passive Intelligent Reflecting Surfaces (IRSs). The neural network significantly outperforms conventional propagation models while remaining computationally tractable. A key finding is that strategically placed IRSs can match the coverage of additional active base stations at a fraction of the energy cost and infrastructure effort, a compelling step toward more sustainable and cost-effective private 6G network deployments in factories and logistics environments.

Ad-Hoc UWB 360° Localization with In-Mask AR Visualization for Emergency Operations
The ComNets team furthermore presented a wearable, infrastructure-free UWB-based 360° localization system for emergency responders operating with respiratory protection. Requiring neither pre-installed infrastructure nor fixed anchor points, the system is well-suited for time-sensitive emergency deployments. Cooperative UWB measurements between body-worn sensor nodes enable real-time determination of each responder's relative position and orientation, delivered into their field of view via an in-mask augmented reality display. Conference visitors were able to test the system interactively, locating hidden UWB nodes using only the distributed UWB localization and AR visualization. 

A Context-Aware Energy Planner for Battery Dimensioning and Technology Selection in 6G-Ready Cellular IoT
The Context-Aware Cellular Energy Planner is an interactive, web-based tool that supports IoT application designers in selecting the right cellular technology and dimensioning batteries appropriately across NB-IoT, RedCap, eRedCap, and 5G NR. The planner combines empirical measurements, simulation-based cellular and energy models, and AI-driven network predictions to optimize for battery life, energy consumption, and sustainability under realistic deployment conditions. A key element is the newly introduced Cellular Eco Score, which goes beyond conventional sustainability metrics by incorporating real-world communication conditions and application-specific battery dimensioning.

The tool is publicly accessible at:
https://energyplanner.kn.e-technik.tu-dortmund.de

Presented Contributions

• N. A. Wagner, C. Wietfeld, "O-RACES: Proactive AI-driven scheduling in Open RAN for 6G-networked humanoid robots," in IEEE INFOCOM Workshops, NetRobiCS Workshop on Networked Robotics and Communication Systems, Tokyo, Japan, May 2026. [pdf] [Video] [Details]
• M. Geis, S. Häger, B. Krluku, C. Wietfeld, “Hybrid Active-Passive Network Design Enabled by Physics-Informed ML-based Propagation Modeling,” in IEEE INFOCOM Workshops, Tokyo, Japan, May 2026. [pdf] [Details]
• P. Jörke, H. Schippers, S. Böcker, C. Wietfeld, “A Context-Aware Energy Planner for Battery Dimensioning and Technology Selection in 6G-Ready Cellular IoT,” in IEEE INFOCOM Workshops, Tokyo, Japan, May 2026. [pdf] [Details]
• F. Schmickmann, P. Jörke, S. Böcker, C. Wietfeld, “Ad-Hoc UWB 360° Localization with In-Mask AR Visualization for Emergency Operations,” in IEEE INFOCOM Workshops, Tokyo, Japan, May 2026. [pdf] [Details]

Acknowledgments
The presented works were supported by the German Federal Ministry of Research, Technology and Space (BMFTR) through the projects CELIDON-PRO (13N16887), 6GEM (16KISK038), GEM-X (16KISS005), ReCoDE (03EI6093B), PANGOLIN Networks (16KIS2357), and the 6GEM+ transfer hub (16KIS2412).