Loading [a11y]/accessibility-menu.js
A Multi-Agent Dueling DQN based Route Selection Scheme for IAB Congestion Controlling | IEEE Conference Publication | IEEE Xplore

A Multi-Agent Dueling DQN based Route Selection Scheme for IAB Congestion Controlling


Abstract:

The millimeter wave (mmWave) band has sufficient bandwidth resources, while it needs a dense base station (BS) deployment for coverage due to higher path loss and lower s...Show More

Abstract:

The millimeter wave (mmWave) band has sufficient bandwidth resources, while it needs a dense base station (BS) deployment for coverage due to higher path loss and lower signal penetration through obstacles. Integrated Access and Backhaul (IAB) becomes the hot spots of research and standardization as a feasible and economical way to achieve dense base station deployment. However, the IAB-node may be affected significantly by congestion problem when transferring data packets, which can result in packet discard and longer transmission delay. This work provides a congestion mitigation scheme based on multi-connectivity (MC) on uplink, which allows congestion nodes utilizing multi-links to improve their backhaul capacity. However, once the IAB-node use multi-links, it will occupy the resources of other IAB-nodes. Therefore, we proposed a Multi-Agent Dueling Deep Q Network (DDQN) based Route Selection Algorithm to help IAB-node decide which links should be selected under different network states. The simulation results present the proposed algorithm outperforms the comparison algorithms in term of packet loss rate and the designed reward.
Date of Conference: 10-13 April 2022
Date Added to IEEE Xplore: 19 May 2022
ISBN Information:

ISSN Information:

Conference Location: Austin, TX, USA

Funding Agency:


Contact IEEE to Subscribe

References

References is not available for this document.