SFG-ROS: A Resource-Aware Framework for Dense Multi-Agent Perception

Constantin Blessing; Dennis Grewe; Elias Geiger; Jakob H\"aringer; Markus Enzweiler

arxiv: 2605.23832 · v1 · pith:ZBVQKTNPnew · submitted 2026-05-22 · 💻 cs.RO

SFG-ROS: A Resource-Aware Framework for Dense Multi-Agent Perception

Constantin Blessing , Elias Geiger , Jakob H\"aringer , Dennis Grewe , Markus Enzweiler This is my paper

Pith reviewed 2026-05-25 03:54 UTC · model grok-4.3

classification 💻 cs.RO

keywords multi-agent perceptionROS 2resource-aware frameworknetwork traffic optimizationdense sensor datacollaborative roboticsCPU scaling reductionschema-driven routing

0 comments

The pith

SFG-ROS bounds multi-agent network traffic to constant order and cuts per-subscriber CPU scaling penalty by 72.3 percent versus standard ROS 2.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Standard ROS 2 implementations encounter network saturation and repeated computational work when dense sensor streams from multiple robots must reach many subscribers. SFG-ROS counters this with schema-driven routing that confines high-frequency traffic locally, an on-demand centralized decoder that performs decompression once and shares results via lightweight IPC, and a container pipeline that adapts to varied robot hardware. Tests on a fleet of wheeled and legged robots carrying LiDAR and stereo cameras show the framework holds network use to O(1) while delivering the measured CPU reduction and preserving low latency. These outcomes matter for any deployment in which robot teams must exchange rich perception data without exhausting bandwidth or processors as the fleet size grows.

Core claim

SFG-ROS isolates intra-agent traffic via a programmatic fully qualified name schema and Fast DDS routing, offloads sensor decompression to a single on-demand pipeline that substitutes lightweight IPC for per-subscriber work, and supplies a hardware-agnostic container layer for zero-touch execution across heterogeneous accelerators; on a mixed wheeled and legged robot fleet the system therefore bounds network traffic to O(1) and reduces the per-subscriber CPU scaling penalty by 72.3 percent relative to unmodified ROS 2 while keeping latency low.

What carries the argument

Schema-driven traffic routing plus on-demand centralized decoding pipeline that replaces redundant decompression with lightweight IPC.

If this is right

Network traffic stays bounded at O(1) as the number of agents and subscribers increases.
CPU overhead for each additional subscriber grows far more slowly than in standard ROS 2.
Low end-to-end latency is retained alongside the traffic and CPU savings.
Heterogeneous robots with different sensors and accelerators can be integrated through the container pipeline without code changes.
Namespace collisions and global network saturation are avoided by design for dynamic fleet deployments.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same separation of local and global traffic plus single-point decoding could be applied to other robot middleware beyond ROS 2.
Energy consumption on battery-powered platforms may improve if the reduced per-subscriber CPU work translates directly to lower power draw.
Verification at fleet sizes beyond the paper's tests would directly test whether the O(1) bound survives real-world wireless contention.
The open-source release permits independent measurement of behavior under sensor types or network conditions not covered in the original evaluation.

Load-bearing premise

The performance measured on the tested fleet of wheeled and legged robots with LiDAR and stereo cameras under the chosen network and hardware conditions extends to general heterogeneous multi-agent deployments without unmeasured trade-offs in reliability or larger-scale behavior.

What would settle it

Repeating the fleet experiments at twenty or more agents on a different network topology and confirming whether network traffic remains O(1) and the 72.3 percent CPU reduction holds.

read the original abstract

Deploying heterogeneous multi-agent robot fleets for collaborative perception requires robust data exchange and scalable software architectures. However, standard ROS 2 implementations often suffer from network saturation, namespace collisions, and severe computational overhead when distributing dense sensor streams across devices. To address these bottlenecks, we present SFG-ROS, a resource-aware multi-agent software framework designed for dynamic fleet deployments. SFG-ROS addresses these challenges through three primary contributions. First, schema-driven traffic routing isolates high-frequency intra-agent traffic from the global network using a programmatic fully qualified name schema and targeted Fast DDS routing. Second, an on-demand centralized decoding pipeline automatically offloads high-bandwidth sensor data decompression, eliminating redundant processing across local consumer nodes. Finally, a hardware-agnostic container pipeline dynamically adapts to heterogeneous accelerators, seamlessly bridging development environments with zero-touch, field-ready execution. We evaluate the framework using a fleet of wheeled and legged robots equipped with LiDAR and stereo depth cameras. Experimental results show SFG-ROS bounds network traffic to $\mathcal{O}(1)$ and, by replacing redundant decompression with lightweight IPC, reduces the per-subscriber CPU scaling penalty by 72.3\% versus standard ROS 2, all while maintaining low latency. Finally, we publish SFG-ROS under a permissive license, available via \href{https://iis-esslingen.github.io/sfg-ros}{iis-esslingen.github.io/sfg-ros}.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

SFG-ROS delivers practical ROS 2 tweaks for multi-agent sensor sharing with measurable CPU gains in small fleets, but the O(1) traffic claim rests on limited scaling evidence.

read the letter

SFG-ROS targets the usual pain points in ROS 2 multi-agent setups: network saturation from dense streams, namespace issues, and repeated decompression work across nodes. The three pieces are a fully qualified name schema plus Fast DDS routing to keep intra-agent traffic local, a central on-demand decoder that does the heavy lifting once, and a container layer that moves between dev and field hardware without changes. That specific mix for ROS 2 looks new based on the abstract and the code release helps others try it directly. The evaluation runs on a mixed fleet of wheeled and legged robots with LiDAR and stereo cameras, which is a reasonable testbed for heterogeneous perception. They report the 72.3% CPU reduction from swapping redundant decompression for lightweight IPC and the O(1) traffic bound while keeping latency low. Those numbers are concrete enough to be worth checking. The O(1) claim is the soft spot. The experiments use a small fleet and the stress-test note is right that there is no explicit scaling sweep or aggregate bandwidth plot versus agent count. Without that, the bound is an extrapolation rather than a demonstrated property for larger deployments. The CPU figure is tied to their exact setup and baselines, so it may not generalize if other factors like reliability under load come into play. This paper is for robotics groups already working with ROS 2 fleets who need engineering relief on bandwidth and CPU. It is the sort of applied systems work that deserves a serious referee to verify the implementation details and experimental controls, even if the scaling part needs tightening. I would send it to review.

Referee Report

2 major / 1 minor

Summary. The manuscript introduces SFG-ROS, a resource-aware framework for dense multi-agent perception in ROS 2. It contributes schema-driven traffic routing via fully qualified names and Fast DDS to isolate high-frequency intra-agent streams from the global network, an on-demand centralized decoding pipeline to eliminate redundant sensor decompression across subscribers, and a hardware-agnostic container pipeline for bridging development and field execution on heterogeneous accelerators. Evaluation on a small fleet of wheeled and legged robots with LiDAR and stereo cameras claims that the framework bounds network traffic to O(1), reduces per-subscriber CPU scaling penalty by 72.3% versus standard ROS 2, and maintains low latency. The implementation is released under a permissive license.

Significance. If the scaling and performance claims hold under broader conditions, the work would provide a practical engineering contribution to multi-robot systems by mitigating network saturation and redundant compute in dense sensor sharing. The open-source release is a positive factor for reproducibility and adoption in the robotics community.

major comments (2)

[Evaluation] Evaluation section: the claim that schema-driven routing bounds global network traffic to O(1) is not demonstrated. Results are reported only for a small heterogeneous fleet (wheeled + legged robots); no scaling sweep with fleet size, asymptotic analysis, or aggregate bandwidth measurements versus number of agents are provided, so the O(1) statement remains an extrapolation rather than a verified property.
[Abstract / Evaluation] Abstract and Evaluation: the quantitative claim of a 72.3% reduction in per-subscriber CPU scaling penalty lacks supporting details on experimental setup, exact baselines, number of subscribers, error bars, or conditions (e.g., sensor rates, network configuration), preventing assessment of whether the data support the stated improvement.

minor comments (1)

[Abstract] The abstract and evaluation could explicitly state the fleet size and number of agents used in the reported experiments to allow readers to contextualize the scaling claims.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major comment point by point below and indicate where revisions will be made to strengthen the evaluation.

read point-by-point responses

Referee: [Evaluation] Evaluation section: the claim that schema-driven routing bounds global network traffic to O(1) is not demonstrated. Results are reported only for a small heterogeneous fleet (wheeled + legged robots); no scaling sweep with fleet size, asymptotic analysis, or aggregate bandwidth measurements versus number of agents are provided, so the O(1) statement remains an extrapolation rather than a verified property.

Authors: The O(1) bound follows from the core design of schema-driven traffic routing: fully qualified names combined with targeted Fast DDS routing isolate all high-frequency intra-agent sensor streams from the global network, so aggregate bandwidth on the shared channel is independent of fleet size. The reported experiments on the heterogeneous wheeled/legged fleet confirm that the isolation mechanism functions as intended under realistic sensor loads. We acknowledge that an explicit empirical scaling sweep across larger fleet sizes is absent. In revision we will add a concise asymptotic analysis of global traffic versus number of agents derived directly from the routing equations, together with any additional aggregate bandwidth plots that can be generated from the existing dataset. revision: partial
Referee: [Abstract / Evaluation] Abstract and Evaluation: the quantitative claim of a 72.3% reduction in per-subscriber CPU scaling penalty lacks supporting details on experimental setup, exact baselines, number of subscribers, error bars, or conditions (e.g., sensor rates, network configuration), preventing assessment of whether the data support the stated improvement.

Authors: The 72.3% figure is obtained by measuring per-subscriber CPU time when replacing redundant sensor decompression with the on-demand centralized decoder plus lightweight IPC, versus unmodified ROS 2. The Evaluation section describes the robot fleet, sensor suite (LiDAR and stereo cameras), and comparison against standard ROS 2. To address the request for greater transparency we will expand the section with explicit values for subscriber count, sensor publication rates, network topology, baseline configurations, and any available error bars or variance measures. revision: yes

Circularity Check

0 steps flagged

No circularity; claims rest on experimental outcomes with no derivations or self-referential fits

full rationale

The manuscript describes a systems framework (schema-driven routing, on-demand decoding, container pipeline) and reports measured outcomes on a small robot fleet (O(1) traffic bound, 72.3% CPU reduction). No equations, parameter fitting, uniqueness theorems, or ansatzes appear. The O(1) statement is presented as an observed experimental property rather than a derived result that reduces to its own inputs. No self-citation chains or renamings of known results are load-bearing. The paper is therefore self-contained against external benchmarks with no circular steps.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on the domain assumption that standard ROS 2 exhibits the stated bottlenecks with dense streams; no free parameters, invented physical entities, or ad-hoc mathematical axioms are introduced in the abstract.

axioms (1)

domain assumption Standard ROS 2 implementations suffer from network saturation, namespace collisions, and severe computational overhead when distributing dense sensor streams across devices.
This premise is stated directly in the abstract as the motivation for the framework.

pith-pipeline@v0.9.0 · 5799 in / 1380 out tokens · 33299 ms · 2026-05-25T03:54:13.032147+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

25 extracted references · 25 canonical work pages

[1]

Procedia Computer Science232, 1972–1981 (2024) https://doi.org/10.1016/j.procs.2024.02.019

Mohr, J., Schmidtke, N., Behrendt, F.: Design of matrix production systems: New demands on factory planning methods. Procedia Computer Science232, 1972–1981 (2024) https://doi.org/10.1016/j.procs.2024.02.019 . 5th International Conference on Industry 4.0 and Smart Manufacturing (ISM 2023)

work page doi:10.1016/j.procs.2024.02.019 1972
[2]

Accessed: 2026-03-27 (2024)

Esslingen University of Applied Sciences: Smart Factory Grids. Accessed: 2026-03-27 (2024). https://www.hs-esslingen.de/en/research/projects/ smart-factory-grids-dfg-forschungsimpuls

work page 2026
[3]

In: Companion Proceedings of the 2024 ACM/IEEE International Conference on Human-Robot Interaction

Malik, A.A., Masood, T., Brem, A.: Intelligent humanoid robots in manufactur- ing. In: Companion Proceedings of the 2024 ACM/IEEE International Conference on Human-Robot Interaction. HRI Companion ’24, pp. 20–27. Association for Computing Machinery, New York, NY, USA (2024). https://doi.org/10.1145/ 3610978.3640765

work page arXiv 2024
[4]

Science Robotics7(66), 6074 (2022) https://doi.org/10.1126/scirobotics.abm6074

Macenski, S., Foote, T., Gerkey, B., Lalancette, C., Woodall, W.: Robot Operat- ing System 2: Design, architecture, and uses in the wild. Science Robotics7(66), 6074 (2022) https://doi.org/10.1126/scirobotics.abm6074

work page doi:10.1126/scirobotics.abm6074 2022
[5]

Open Robotics

ROS Core Team: 2023-09 ROS 2 RMW Alternate. Open Robotics. Accessed: 2026-03-27 (2023). https://discourse.openrobotics.org/uploads/short-url/ o9ihvSjCwB8LkzRklpKdeesRTDi.pdf

work page 2023
[6]

Accessed: 2026-03-25 (2021)

Open Robotics: Open-RMF: Open Robotics Middleware Framework. Accessed: 2026-03-25 (2021). https://www.open-rmf.org/

work page 2026
[7]

In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp

Sikand, K.S., Zartman, L., Rabiee, S., Biswas, J.: Robofleet: Open source com- munication and management for fleets of autonomous robots. In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 406–412 (2021). IEEE

work page 2021
[8]

In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp

Ichnowski, J., Chen, K., Dharmarajan, K., Adebola, S., Danielczuk, M., Mayoral- Vilches, V., Jha, N., Zhan, H., LLontop, E., Xu, D.,et al.: FogROS2: An adaptive 18 platform for cloud and fog robotics using ROS 2. In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp. 5493–5500 (2023). IEEE

work page 2023
[9]

Accessed: 2026-03-31 (2026)

Open Robotics: Deployment Guidelines — ROS 2 Documentation: Humble documentation. Accessed: 2026-03-31 (2026). https://docs.ros.org/en/humble/ Tutorials/Advanced/Security/Deployment-Guidelines.html

work page 2026
[10]

Processes 14(5) (2026) https://doi.org/10.3390/pr14050804

Lim, J.M., Kim, K.W., Choi, B.W., Delgado, R.: A docker-enabled real-time framework for robotic applications in heterogeneous ros 2 environments. Processes 14(5) (2026) https://doi.org/10.3390/pr14050804

work page doi:10.3390/pr14050804 2026
[11]

In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp

Zhang, Y., Wurll, C., Hein, B.: KubeROS: A unified platform for automated and scalable deployment of ROS 2-based multi-robot applications. In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp. 9097–9103 (2023). IEEE

work page 2023
[12]

Accessed: 2026-03-25 (2021)

NVIDIA Corporation: NVIDIA Isaac ROS: Hardware-Accelerated Packages for ROS 2 Developers. Accessed: 2026-03-25 (2021). https://developer.nvidia.com/ isaac/ros

work page 2026
[13]

Versions: 2.6.2 (ROS 2), 3.2.2 (DDS Router) (2026)

eProsima: Fast DDS. Versions: 2.6.2 (ROS 2), 3.2.2 (DDS Router) (2026). https: //github.com/eProsima/Fast-DDS

work page 2026
[14]

Version: 3.2.0 (2026)

eProsima: DDS Router. Version: 3.2.0 (2026). https://github.com/eProsima/ DDS-Router

work page 2026
[15]

Version: 3.0.2

Pfrommer, B.: ROS 2 Image Transport for FFmpeg/Libav. Version: 3.0.2. https: //github.com/ros-misc-utilities/ffmpeg image transport

work page
[16]

Signal Processing: Image Communication24(1), 73–88 (2009) https://doi

Merkle, P., Morvan, Y., Smolic, A., Farin, D., M¨ uller, K., de With, P.H.N., Wiegand, T.: The effects of multiview depth video compression on multiview ren- dering. Signal Processing: Image Communication24(1), 73–88 (2009) https://doi. org/10.1016/j.image.2008.10.010 . Special issue on advances in three-dimensional television and video

work page doi:10.1016/j.image.2008.10.010 2009
[17]

IEEE Trans- actions on Circuits and Systems for Video Technology26(1), 35–49 (2016) https://doi.org/10.1109/TCSVT.2015.2477935

Tech, G., Chen, Y., M¨ uller, K., Ohm, J.-R., Vetro, A., Wang, Y.-K.: Overview of the multiview and 3D extensions of high efficiency video coding. IEEE Trans- actions on Circuits and Systems for Video Technology26(1), 35–49 (2016) https://doi.org/10.1109/TCSVT.2015.2477935

work page doi:10.1109/tcsvt.2015.2477935 2016
[18]

In: Proceedings of the 2017 ACM International Conference on Interactive Surfaces and Spaces

Wilson, A.D.: Fast lossless depth image compression. In: Proceedings of the 2017 ACM International Conference on Interactive Surfaces and Spaces. ISS ’17, pp. 100–105. Association for Computing Machinery, New York, NY, USA (2017). https://doi.org/10.1145/3132272.3134144

work page doi:10.1145/3132272.3134144 2017
[19]

Telecommunication Systems60(2), 337–345 (2015) 19

Kam, H.R., Lee, S.-H., Park, T., Kim, C.-H.: RViz: a toolkit for real domain data visualization. Telecommunication Systems60(2), 337–345 (2015) 19

work page 2015
[20]

Rerun Technologies AB, Online

Rerun Development Team: Rerun: A Visualization SDK for Multimodal Data. Rerun Technologies AB, Online. Version: 0.25.1 (2024). https://www.rerun.io

work page 2024
[21]

Accessed: 2026- 3-18

Vicon Motion Systems: Vicon Vero Camera for Motion Capture. Accessed: 2026- 3-18. https://www.vicon.com/hardware/cameras/vero/

work page 2026
[22]

Accessed: 2026-3-18

Vicon Motion Systems: Vicon Tracker 4.4 documentation. Accessed: 2026-3-18. https://help.vicon.com/space/Tracker44

work page 2026
[23]

Version: 1.0.2

OPT4SMART group: Vicon Receiver for ROS 2. Version: 1.0.2. https://github. com/OPT4SMART/ros2-vicon-receiver

work page
[24]

ACM SIGGRAPH Computer Graphics22(4), 199–204 (1988) Author Biographies Constantin Blessingreceived his M.Sc

Barry, P.J., Goldman, R.N.: A recursive evaluation algorithm for a class of Catmull-Rom splines. ACM SIGGRAPH Computer Graphics22(4), 199–204 (1988) Author Biographies Constantin Blessingreceived his M.Sc. degree in applied computer science from the Esslingen University of Applied Sciences, Esslingen am Neckar, Germany, in 2023, where he is currently work...

work page 1988
[25]

Since September 2024, he has been a Full Professor of Distributed Systems and Software Testing with the Esslingen University of Applied Sciences, Esslin- gen am Neckar, Germany

From 2019 to 2023, he was a Research Engineer with Robert Bosch GmbH. Since September 2024, he has been a Full Professor of Distributed Systems and Software Testing with the Esslingen University of Applied Sciences, Esslin- gen am Neckar, Germany. His research interests include network protocols and architectures to support comput- ing on network infrastr...

work page 2019

[1] [1]

Procedia Computer Science232, 1972–1981 (2024) https://doi.org/10.1016/j.procs.2024.02.019

Mohr, J., Schmidtke, N., Behrendt, F.: Design of matrix production systems: New demands on factory planning methods. Procedia Computer Science232, 1972–1981 (2024) https://doi.org/10.1016/j.procs.2024.02.019 . 5th International Conference on Industry 4.0 and Smart Manufacturing (ISM 2023)

work page doi:10.1016/j.procs.2024.02.019 1972

[2] [2]

Accessed: 2026-03-27 (2024)

Esslingen University of Applied Sciences: Smart Factory Grids. Accessed: 2026-03-27 (2024). https://www.hs-esslingen.de/en/research/projects/ smart-factory-grids-dfg-forschungsimpuls

work page 2026

[3] [3]

In: Companion Proceedings of the 2024 ACM/IEEE International Conference on Human-Robot Interaction

Malik, A.A., Masood, T., Brem, A.: Intelligent humanoid robots in manufactur- ing. In: Companion Proceedings of the 2024 ACM/IEEE International Conference on Human-Robot Interaction. HRI Companion ’24, pp. 20–27. Association for Computing Machinery, New York, NY, USA (2024). https://doi.org/10.1145/ 3610978.3640765

work page arXiv 2024

[4] [4]

Science Robotics7(66), 6074 (2022) https://doi.org/10.1126/scirobotics.abm6074

Macenski, S., Foote, T., Gerkey, B., Lalancette, C., Woodall, W.: Robot Operat- ing System 2: Design, architecture, and uses in the wild. Science Robotics7(66), 6074 (2022) https://doi.org/10.1126/scirobotics.abm6074

work page doi:10.1126/scirobotics.abm6074 2022

[5] [5]

Open Robotics

ROS Core Team: 2023-09 ROS 2 RMW Alternate. Open Robotics. Accessed: 2026-03-27 (2023). https://discourse.openrobotics.org/uploads/short-url/ o9ihvSjCwB8LkzRklpKdeesRTDi.pdf

work page 2023

[6] [6]

Accessed: 2026-03-25 (2021)

Open Robotics: Open-RMF: Open Robotics Middleware Framework. Accessed: 2026-03-25 (2021). https://www.open-rmf.org/

work page 2026

[7] [7]

In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp

Sikand, K.S., Zartman, L., Rabiee, S., Biswas, J.: Robofleet: Open source com- munication and management for fleets of autonomous robots. In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 406–412 (2021). IEEE

work page 2021

[8] [8]

In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp

Ichnowski, J., Chen, K., Dharmarajan, K., Adebola, S., Danielczuk, M., Mayoral- Vilches, V., Jha, N., Zhan, H., LLontop, E., Xu, D.,et al.: FogROS2: An adaptive 18 platform for cloud and fog robotics using ROS 2. In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp. 5493–5500 (2023). IEEE

work page 2023

[9] [9]

Accessed: 2026-03-31 (2026)

Open Robotics: Deployment Guidelines — ROS 2 Documentation: Humble documentation. Accessed: 2026-03-31 (2026). https://docs.ros.org/en/humble/ Tutorials/Advanced/Security/Deployment-Guidelines.html

work page 2026

[10] [10]

Processes 14(5) (2026) https://doi.org/10.3390/pr14050804

Lim, J.M., Kim, K.W., Choi, B.W., Delgado, R.: A docker-enabled real-time framework for robotic applications in heterogeneous ros 2 environments. Processes 14(5) (2026) https://doi.org/10.3390/pr14050804

work page doi:10.3390/pr14050804 2026

[11] [11]

In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp

Zhang, Y., Wurll, C., Hein, B.: KubeROS: A unified platform for automated and scalable deployment of ROS 2-based multi-robot applications. In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp. 9097–9103 (2023). IEEE

work page 2023

[12] [12]

Accessed: 2026-03-25 (2021)

NVIDIA Corporation: NVIDIA Isaac ROS: Hardware-Accelerated Packages for ROS 2 Developers. Accessed: 2026-03-25 (2021). https://developer.nvidia.com/ isaac/ros

work page 2026

[13] [13]

Versions: 2.6.2 (ROS 2), 3.2.2 (DDS Router) (2026)

eProsima: Fast DDS. Versions: 2.6.2 (ROS 2), 3.2.2 (DDS Router) (2026). https: //github.com/eProsima/Fast-DDS

work page 2026

[14] [14]

Version: 3.2.0 (2026)

eProsima: DDS Router. Version: 3.2.0 (2026). https://github.com/eProsima/ DDS-Router

work page 2026

[15] [15]

Version: 3.0.2

Pfrommer, B.: ROS 2 Image Transport for FFmpeg/Libav. Version: 3.0.2. https: //github.com/ros-misc-utilities/ffmpeg image transport

work page

[16] [16]

Signal Processing: Image Communication24(1), 73–88 (2009) https://doi

Merkle, P., Morvan, Y., Smolic, A., Farin, D., M¨ uller, K., de With, P.H.N., Wiegand, T.: The effects of multiview depth video compression on multiview ren- dering. Signal Processing: Image Communication24(1), 73–88 (2009) https://doi. org/10.1016/j.image.2008.10.010 . Special issue on advances in three-dimensional television and video

work page doi:10.1016/j.image.2008.10.010 2009

[17] [17]

IEEE Trans- actions on Circuits and Systems for Video Technology26(1), 35–49 (2016) https://doi.org/10.1109/TCSVT.2015.2477935

Tech, G., Chen, Y., M¨ uller, K., Ohm, J.-R., Vetro, A., Wang, Y.-K.: Overview of the multiview and 3D extensions of high efficiency video coding. IEEE Trans- actions on Circuits and Systems for Video Technology26(1), 35–49 (2016) https://doi.org/10.1109/TCSVT.2015.2477935

work page doi:10.1109/tcsvt.2015.2477935 2016

[18] [18]

In: Proceedings of the 2017 ACM International Conference on Interactive Surfaces and Spaces

Wilson, A.D.: Fast lossless depth image compression. In: Proceedings of the 2017 ACM International Conference on Interactive Surfaces and Spaces. ISS ’17, pp. 100–105. Association for Computing Machinery, New York, NY, USA (2017). https://doi.org/10.1145/3132272.3134144

work page doi:10.1145/3132272.3134144 2017

[19] [19]

Telecommunication Systems60(2), 337–345 (2015) 19

Kam, H.R., Lee, S.-H., Park, T., Kim, C.-H.: RViz: a toolkit for real domain data visualization. Telecommunication Systems60(2), 337–345 (2015) 19

work page 2015

[20] [20]

Rerun Technologies AB, Online

Rerun Development Team: Rerun: A Visualization SDK for Multimodal Data. Rerun Technologies AB, Online. Version: 0.25.1 (2024). https://www.rerun.io

work page 2024

[21] [21]

Accessed: 2026- 3-18

Vicon Motion Systems: Vicon Vero Camera for Motion Capture. Accessed: 2026- 3-18. https://www.vicon.com/hardware/cameras/vero/

work page 2026

[22] [22]

Accessed: 2026-3-18

Vicon Motion Systems: Vicon Tracker 4.4 documentation. Accessed: 2026-3-18. https://help.vicon.com/space/Tracker44

work page 2026

[23] [23]

Version: 1.0.2

OPT4SMART group: Vicon Receiver for ROS 2. Version: 1.0.2. https://github. com/OPT4SMART/ros2-vicon-receiver

work page

[24] [24]

ACM SIGGRAPH Computer Graphics22(4), 199–204 (1988) Author Biographies Constantin Blessingreceived his M.Sc

Barry, P.J., Goldman, R.N.: A recursive evaluation algorithm for a class of Catmull-Rom splines. ACM SIGGRAPH Computer Graphics22(4), 199–204 (1988) Author Biographies Constantin Blessingreceived his M.Sc. degree in applied computer science from the Esslingen University of Applied Sciences, Esslingen am Neckar, Germany, in 2023, where he is currently work...

work page 1988

[25] [25]

Since September 2024, he has been a Full Professor of Distributed Systems and Software Testing with the Esslingen University of Applied Sciences, Esslin- gen am Neckar, Germany

From 2019 to 2023, he was a Research Engineer with Robert Bosch GmbH. Since September 2024, he has been a Full Professor of Distributed Systems and Software Testing with the Esslingen University of Applied Sciences, Esslin- gen am Neckar, Germany. His research interests include network protocols and architectures to support comput- ing on network infrastr...

work page 2019