Constarium
← Search

Data · dataset · 2026

Dataset for the Paper "O-RACES: Proactive AI-driven Scheduling in Open RAN for 6G-Networked Humanoid Robots"

Listed in TUDOdata

O-RACES: Proactive AI-driven Scheduling in Open RAN for 6G-Networked Humanoid Robots This repository contains the O-RACES measurements dataset.

Description

It is supplementary material to the paper “O-RACES: Proactive AI-driven Scheduling in Open RAN for 6G-Networked Humanoid Robots” accepted for IEEE INFOCOM Workshops in May 2026. The dataset includes CSV logs from humanoid robot locomotion experiments (robot observations/actions) and accompanying Edge-Cloud controller and Open RAN network metrics from offloading the locomotion control wirelessly to the edge cloud.

While the (up to 16) humanoid robots are simulated in real-time using NVIDIA Isaac Sim, the network used is a real-world Open RAN testbed. State data from the robot are sent via one modem per each robot to the Edge Cloud, where the Reinforcement Learning (RL)-trained locomotion control policy is deployed. The controller then sends actions via the real-world Open RAN to the robot, which executes them in the simulator.

Read the rest (4 more)

The dataset includes experimental data for three different uplink schedulers: a reactive scheduling baseline, a static pre-scheduling baseline, and the proposed O-RACES proactive scheduler. A demo video of the experiments is available at tiny.cc/O-RACES-Humanoid. If you use this dataset, please cite the corresponding paper.

The pre-print version of the paper is available under: Pre-Print Paper. Directory structure Data is grouped by scenario (used scheduler) and by the number of parallel environments (robots) used during the experiment: reactive/<N>/ (Conventional Reactive Uplink Slicing Scheduling) static/<N>/ (Static Uplink Pre-Scheduling) predictive/<N>/ (Proposed O-RACES Scheduler) where <N> is one of 1, 2, 4, 8, 16. Each directory contains 16 runs run01, run02, ... , run16 Description of Files per run Each run consists of three CSV files: File: runXX_robot.csv Description: Robot log (state, commands, actions, torques, observations) Time columns: timestamp_ms, sim_time_s File: runXX_controller.csv Description: Controller-side metrics at the AI Edge Cloud Time columns: timestamp_ms File: runXX_ran.csv Description: Real-World Open RAN network metrics Time columns: timestamp_ms Timebase: - timestamp_ms is aligned across files and starts at 0 for each run. - sim_time_s is the robot simulation time and starts at 0 for each run.

File Details: Information about the specific files is included in the README.md file in the dataset root directory. Support If you have any issues with the supplied data, please feel free to contact one of the article's authors. Acknowledgments This work has been funded by the Federal Ministry of Research, Technology and Space (BMFTR) via the 6GEM research hub and the 6GEM+ transfer hub under funding references 16KISK038 and 16KIS2412.

How to cite Please cite this paper if you use the data: N. A. Wagner, C. Wietfeld, “O-RACES: Proactive AI-driven Scheduling in Open RAN for 6G-Networked Humanoid Robots,” IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), May 2026. @inproceedings{wagner2026oraces, author = {Niklas A. Wagner and Christian Wietfeld}, title = {{O-RACES}: Proactive {AI}-driven Scheduling in {Open RAN} for {6G}-Networked Humanoid Robots}, booktitle = {IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)}, location = {Tokyo, Japan}, year = {2026} }

Links

Topics

Inferred from text
Artificial intelligence 73% · Simulation 75% · Video 75%
Provenance · 1 source records, 16 field assertions
SourceKeyLast seenRaw
TUDOdatadoi:10.17877/TUDODATA-2026-MLFQ463Q9 d agoJSON v1
FieldAssertionExtractorEvidence
concepts[field].anzsrc:group:4602enrichment · data tu dortmund detaxonomy-embedding@1.0.0title+keywords+description (73%)
concepts[field].dataverse_subject:computer-and-information-sciencesource · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0/subjects
concepts[field].dataverse_subject:engineeringsource · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0/subjects
concepts[field].local:field:chemistrymapping · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0/subjects
concepts[field].local:field:engineeringmapping · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0/subjects
concepts[field].local:field:humanitiesmapping · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0/subjects
concepts[field].local:field:mathematics-statisticsmapping · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0/subjects
concepts[field].local:field:social-sciencemapping · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0/subjects
concepts[method].local:method:simulationenrichment · data tu dortmund dekeyword-concept-rules@1.0.0title+description (75%)
concepts[modality].local:modality:videoenrichment · data tu dortmund dekeyword-concept-rules@1.0.0title+description (75%)
created_datesource · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0
descriptionsource · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0/description
publication_datesource · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0
titlesource · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0/name
updated_datesource · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0
version_labelsource · data tu dortmund deconnector:data_tu_dortmund_de@1.0.0