Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • apa.csl
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Investigating Machine Learning Training Strategies for 5G Latency Prediction Across Paths and Temporal Context
Universita degli Studi di Napoli Federico II, Italy.
Universita degli Studi di Napoli Federico II, Italy.
Karlstad University, Faculty of Health, Science and Technology (starting 2013), Department of Mathematics and Computer Science (from 2013).ORCID iD: 0000-0003-0611-5637
Universita degli Studi di Napoli Federico II, Italy.
Show others and affiliations
2026 (English)In: IEEE International Symposium on Measurement in Networking and Communications (MNC), Institute of Electrical and Electronics Engineers (IEEE), 2026Conference paper, Published paper (Refereed)
Abstract [en]

Predicting latency performance is crucial for many delay-sensitive applications running nowadays on 5G systems. As a matter of fact, anticipating latency degradation can help improve system reliability by triggering network configuration and policy adjustments, e.g., in terms of traffic and user scheduling. Machine Learning (ML) offers an effective approach for this prediction task and is considered a key enabler for future 6G networks standardized by the 3rd Generation Partnership Project (3GPP). In this work, we analyze whether it is possible to reliably predict latency performance by using ML approaches, and investigate how to perform such a task. To do so, we leverage extensive measurements collected in a real-world 5G network deployed in the Karlstad University campus (Karlstad, Sweden). The dataset includes latency performance, in terms of Round-Trip Time (RTT) and radio coverage logs, collected across mobility scenarios, server locations, and application traffic patterns. We thus leverage this heterogeneous dataset to investigate different training strategies, including mixing data across on-campus paths/locations, separating them, and incorporating temporal features such as RTT lags. We also explore the use of measurements from spatially adjacent paths. Results show that accurate prediction (around 85%) can be achieved, e.g., by training ML models on a subset of paths and testing them on other subsets, with clear accuracy gains from both temporal and spatial information. 

Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers (IEEE), 2026.
Keywords [en]
Learning algorithms, Learning systems, Machine learning, Mobile telecommunication systems, 5g, Latency, Latency performance, Latency predictions, Machine-learning, Measurement campaign, Round-trip-time, Smart phones, Smartphone-based, Training strategy, Forecasting
National Category
Computer Sciences Computer Systems Communication Systems
Research subject
Computer Science
Identifiers
URN: urn:nbn:se:kau:diva-112659DOI: 10.1109/MNC69143.2026.11680705Scopus ID: 2-s2.0-105051678774ISBN: 979-8-3315-6214-4 (electronic)OAI: oai:DiVA.org:kau-112659DiVA, id: diva2:2114134
Conference
IEEE International Symposium on Measurement in Networking and Communications (MNC), Naples, Italy, July 13-16, 2026.
Funder
Knowledge FoundationEU, Horizon Europe, 101293242Available from: 2026-10-30 Created: 2026-10-30 Last updated: 2026-10-30Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full textScopus

Authority records

Caso, GiuseppeBrunstrom, Anna

Search in DiVA

By author/editor
Caso, GiuseppeBrunstrom, Anna
By organisation
Department of Mathematics and Computer Science (from 2013)
Computer SciencesComputer SystemsCommunication Systems

Search outside of DiVA

GoogleGoogle Scholar

doi
isbn
urn-nbn

Altmetric score

doi
isbn
urn-nbn
Total: 1 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • apa.csl
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf