University of Twente Student Theses


AI Enabled Slice Resource Management for Improved QoS in WiFi-based IoT Networks

Harten, R.H. van (2023) AI Enabled Slice Resource Management for Improved QoS in WiFi-based IoT Networks.

[img] PDF
Abstract:As new IoT services are being developed, the number of connected devices with different QoS requirements in many WiFi networks is ever-increasing. The current standard of using different access categories is often not well suited to provide the wide variety of QoS requirements of the different connected devices. Network slicing, where network airtime is divided into segments called slices, is a potent technology to meet diverse QoS requirements in IoT networks. Traffic flows with different QoS requirements are assigned to slices and their resources are managed to prioritize flows over one another and subsequently meet their requirements. Proposed solutions have considered a limited number of slices, however, and QoS diversity in IoT requires eight or more slices to efficiently meet QoS requirements. Managing the resources of a larger number of slices is a complex task, but AI can be used to handle this complexity. Therefore, we have implemented a DRL agent using DDPG that can divide the airtime between the slices in a simulated IoT network with eight slices and manages to nearly provide the necessary throughput requirements in a network where resources are limited.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science BSc (56964)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page