Network Digital Twin for 5G-Enabled Mobile Robots
Luis Roda-Sanchez 1 , Lanfranco Zanzi 2 , Xi Li 2 , Guillem Garí 3 , Xavier Costa-Pérez 4 2 5 · 2025
Read Network Digital Twin for 5G-Enabled Mobile Robots by Luis Roda-Sanchez 1 , Lanfranco Zanzi 2 , Xi Li 2 , Guillem Garí 3 , Xavier Costa-Pérez 4 2 5 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-02253.txt.
Introduction
Luis Roda Sanchez, Lanfranco Zanzi, and their co-authors build a digital twin of a 5G network for mobile robots. Their 2025 paper notes that private 5G gives robots low delay, high speed links, but network quality shifts fast in busy halls and yards. Robots need a live picture of that quality to navigate and work reliably.
Their framework feeds real time data collected by the robots themselves into the twin. The twin then supports monitoring, control, and tuning of robot operations as conditions change. A pipeline carries robot measurements into the model, and the paper shows how the twin evolves with fresh data. The approach turns the fleet into its own network sensor, closing the loop between coverage maps and robot behavior.
Catalog record
| CREATOR | Luis Roda-Sanchez 1 , Lanfranco Zanzi 2 , Xi Li 2 , Guillem Garí 3 , Xavier Costa-Pérez 4 2 5 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-02253
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Network Digital Twin for 5G-Enabled Mobile Robots Luis Roda-Sanchez 1 , Lanfranco Zanzi 2 , Xi Li 2 , Guillem Garí 3 , Xavier Costa-Pérez 4 2 5 Abstract The maturity and commercial roll-out of 5G networks and its deployment for private networks makes 5G a key enabler for various vertical industries and applications, including robotics. Providing ultra-low latency, high data rates, and ubiquitous coverage and wireless connectivity, 5G fully unlocks the potential of robot autonomy and boosts emerging robotic applications, particularly in the domain of autonomous mobile robots. Ensuring seamless, efficient, and reliable navigation and operation of robots within a 5G network requires a clear understanding of the expected network quality in the deployment environment. However, obtaining real-time insights into network conditions, particularly in highly dynamic environments, presents a significant and practical challenge. In this paper, we present a novel framework for building a Network Digital Twin (NDT) using real-time data collected by robots. This framework provides a comprehensive solution for monitoring, controlling, and optimizing robotic operations in dynamic network environments. We develop a pipeline integrating robotic data into the NDT, demonstrating its evolution with real-world robotic traces. We evaluate its performances in radio-aware navigation use case, highlighting its potential to enhance energy efficiency and reliability for 5G-enabled robotic operations. 4 i2CAT Foundation, Barcelona, Spain. 5 ICREA, Barcelona, Spain. I Introduction The field of robotics has experienced notable progress in the realm of autonomous mobile robots, now equipped with wireless technologies like WiFi and 5G, enabling remote operation and task offloading [ 1 ] . These robots hold the promise of transforming various industries by autonomously and efficiently performing tasks in dynamic environments. Autonomous robots rely on onboard sensors to build environmental knowledge of their proximity by collecting sensor data along their navigation trajectory, processing this information to build a comprehensive understanding of their environment [ 2 ] . This results in expensive and time-consuming navigation efforts when a global network quality of the whole environment is needed. At the same time, varying radio quality can hinder the data acquisition process, which in turn affects the robot’s ability to make timely decisions and maintain precise positioning. To tackle this challenge, Network Digital Twin (NDT) technology emerges as a promising solution, offering the ability to estimate the expected quality of a physical network. It does so by leveraging a combination of measured or estimated data collected from both the network infrastructure and the robots or other User Equipment (UEs). Such NDT is not merely a static model but a dynamic and data-driven one that mirrors the behaviour, components, and interactions of the physical network in near real-time [ 3 ] . This would allow mobile robots to make more informed decisions and adapt their autonomous operations over run time. For instance, mobile robots would benefit from optimal radio link conditions for computational offloading, helping conserve energy that would otherwise be used to transmit data under poor channel conditions. However, the research of NDT technology in the field of communication networks is still in its early phase. Most of the NDT works rely on collected or simulated network data, and take a system-wide approach from a network operator’s point of view by assuming full knowledge of the network infrastructure. The drawback of such approach is not only costly (in terms of model complexity and large amount of data to collect), but also slow to adapt to network changes, which is incapable of capturing dynamic changes over run time. Conversely, we aim to build an NDT from the UE (or a robot) perspective, which is not aware of the deployment environment and of the available network infrastructure. Thus, in this paper we propose a novel framework to build an online NDT from a robot when exploring an unknown area and its corresponding network environment through Robot Operating System (ROS), for estimation of the radio quality of 5G networks. Based on the proposed framework, we design a pipeline to exploit ROS-based information when creating and updating the NDT via a dedicated API. We explore the setup of an NDT at run time, including methods for data collection, modelling techniques, and their integration with robotic control algorithms. To validate the proposed NDT framework, we exploit robotic traces collected in an operational environment to showcase the evolution of an NDT model and its capabilities to adapt to variable network scenarios. And finally we show an example of using NDT for radio-aware navigation and highlight its benefits in improving energy efficiency and reliability for 5G robot operations. II Related Works By leveraging the capabilities of NDT, digital replicas of networked environments can be created to offer unparalleled insights, predictive capabilities, and opportunities for optimization. 5G’s low latency, high bandwidth, and massive device connectivity provide the foundation for real-time communication and control, enabling robots to readily respond to unexpected dynamics in the deployment environment. In this context, communication-aware motion planning in robotic networks is becoming of key importance in the robotic domain, due to the impelling need of offloading heavy computation tasks to edge platforms. The authors of [ 4 ] investigate radio-aware semantic map creation along the exploration of unknown environments, and provide simulation-based results and a ROS package to deal with such scenarios. Similarly, [ 5 ] investigate joint optimization of 5G and robotic domain as a way to improve resource allocation and energy consumption KPIs. In the domain of wireless channel estimation within robotic networks, the authors of [ 6 ] propose a framework for estimating spatial variations in a wireless channel within a robotic network. They introduce a multiscale probabilistic model to characterize the wireless channel and develop an estimator based on this model. In the area of NDT, most of the works in the literature consider holistic approaches by assuming full knowledge of the network infrastructure and environment, for example, the authors of [ 7 ] propose a knowledge graph-based construction method for NDTs, exploiting a graph representation to model the system dynamics. Altogether, this highlights the ongoing innovation and evolution in both the fields of robotics and communication to build an NDT, especially from a robot perspective, which needs to discover its environment and available network infrastructure. Key challenges in building an NDT for robots stand on the model accuracy and real-time data acquisition. In this paper we introduce an online NDT model for robots exploring unknown areas, utilizing ROS-based real-time data to create and update NDT efficiently. II-A Architecture Framework To build such an NDT to integrate with robotic data and applications and use it for joint network and robotic control, we design the architecture framework depicted in Fig. 1 . It includes the Physical System, Digital Twin Layer, and Digital Twin Application layer. The Digital Twin Layer is the digital representation of the Physical System, which are interconnected with each other in real-time to realize the closed loop for the control of the real work system. The digital representation is composed of Network Modeling and Physical Environment Modeling. The physical environment modeling creates the digital modelling of the physical environment, including the space, topology, objects, devices, humans, etc. The Network Modeling allows for building the virtual representation of the network behavior, as well as estimating the expected network performance