Scope
The Key Performance Indicators identified for 5G wireless networks impose the application of comprehensive, sophisticated and energy-efficient algorithms and solutions at both radio access and core network, but also on data centres and storage. It is widely understood that the reliable and immediate access to accurate data defining – in a broad sense – the whole communication context may significantly improve the overall performance of the system. In the context of future wireless networks, acquiring large amounts of data to harvest correlations and statistical probabilities are envisaged to enable proactive decisions and thereby improving network performance and efficiency. Thus, application of machine learning algorithms including artificial intelligence based solutions may be necessary. Although the data structures can be structured in many ways, correlating the information with geolocation is a promising concept providing both efficiency & visualization. This can be manifested as Radio Environment Maps (REM) capturing statistics on channel quality, throughput and link reliability etc. Traffic maps (e.g. distribution of requested or served traffic, traffic patterns etc.) and mobility, migration and trajectory information can enable improved short- and long term proactive resource management in the network. Implementation of historical knowledge related to the users and cells, migration patterns of users etc. may be applied for better caching of data content, fog computing and MEC. Finally, rich knowledge of user behaviour can be utilized for improving energy efficiency in future networks as, for example, selected sectors (or individual carriers) of base stations can be switched off based on historical data and traffic prediction maps. In a nutshell, an access to the accurate and rich information defining the communication context (known as context information) can lead to significant improvement of various performance metrics and predictive maintenance.
However, the more data required by various algorithms, the higher the load of the control plane and the stronger requirements on backhaul part of the system (which includes maintenance and access to data centers and storage), as well as higher energy costs. Furthermore, in dense and heterogonous networks the amount of prospective context information may be huge, what in consequence leads to the issue of big-data processing with high geographical granularity. Moreover, application of artificial intelligence algorithms may be necessary for proper inference and reasoning based on available rich context information.
This workshop deals with the aspects of big data processing and application of artificial intelligence solutions in future wireless networks for better resource management and for achieving higher energy efficiency. It will try to answer the following key issues:
- How to initialize various types of data (e.g., radio environment maps, radio service maps, RSMs) to achieve a decent efficiency already from start
- How to use REMs/RSMs and other databases to adapt resource availability (enable/disable, bandwidth, power settings etc.) for better network sharing, content delivery and energy efficiency
- How to steer traffic in the most efficient way given the traffic characteristics, network load, radio conditions and available resources
- How to characterize users behavior (traffic types and intensity, mobility patterns, etc.) and their channels (measurements, CSI reports etc.),
- How to characterize the network resources in terms of what is important to model and keep track of to achieve the end goals above
- How to orchestrate alternative connectivity solutions for enhancing user experience, network performance and/or energy efficiency
- How to visualize the data structures to ensure that network is operating according to policies
Key Topics
This workshop focus on how to use big data processing and artificial intelligence in future wireless networks to improve network performance, radio resource utilization and energy efficiency, while delivering expected QoE. The topics covered by this workshop include, but are not limited to:
- Application of radio environment and service maps for resource management and energy efficiency
- Big data processing at 5G RAN and core for network performance and energy efficiency
- Application of artificial intelligence algorithms for big data analysis in 5G networks
- Application of rich context information for fog- and cloud computing, and MEC
- Big data delivery and application of AI in 5G
- Context Aware communications boosted by artificial intelligence
- Database supported resource and interference management
- Data storage, processing, analysis for RAN
- Big data analysis and AI for SON and network slicing
- Big data for predictive maintenance
- Orchestration of future networks
- Visualization aspects of data
Workshop Organizers
- Dr. Adrian Kliks, Poznan University of Technology, Poland (contact person, e-mail: adrian.kliks@put.poznan.pl)
- Marcin Dryjanski, Huawei Technologies, Sweden
- Dr. Magnus Isaksson Huawei Technologies, Sweden
- Dr. Azeddine Gati, Orange Labs, Paris, France
Important dates
- Submission deadline:
November 2, 2017extended November 12, 2017 - Notification of Acceptance: December 15, 2017
- Camera-Ready Submission: January 21, 2018
Submission policy
This workshop only accepts novel, previously unpublished papers. All submissions should be written in English with a maximum paper length of five (5) printed pages (IEEE Trans. Conf format) including figures without incurring additional page charges (maximum 1 additional page with over-length page charge if accepted). All papers will be subject to a rigorous peer-review process. all accepted papers will be published in an IEEE Explore database.
While submission please follow the general rules provided here: LINK
Submission has to be done via the EDAS system - DIRECT LINK
PDF version
Call-for-papers in pdf format - LINK
