Intelligent Partial-Sensing-Based Autonomous Resource Allocation for NR V2X
- Authors
- Kim, Taehyoung; Kim, Younsun; Jung, Minchae; Son, Hyukmin
- Issue Date
- Jan-2024
- Publisher
- IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
- Keywords
- Blind decoding; channel-type classification; con-trol channel; convolutional neural network; new radio (NR); power consumption; resource allocation (RA) mode 2; sidelink (SL); vehicle-to-everything (V2X)
- Citation
- IEEE INTERNET OF THINGS JOURNAL, v.11, no.2, pp 3144 - 3160
- Pages
- 17
- Journal Title
- IEEE INTERNET OF THINGS JOURNAL
- Volume
- 11
- Number
- 2
- Start Page
- 3144
- End Page
- 3160
- URI
- https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/90820
- DOI
- 10.1109/JIOT.2023.3295024
- ISSN
- 2327-4662
- Abstract
- Since its introduction for long term evolution (LTE), vehicle-to-everything (V2X) communications have evolved to the recent new radio (NR)-based V2X with enhanced reliability, latency, capacity, and flexibility. One of the key features of NR V2X is the enabling of sidelink (SL) communications between user equipments (UEs) without any assistance from a base station (i.e., support of the out-of-coverage scenario). To this end, NR V2X supports SL resource allocation (RA) mode 2, in which a UE autonomously determines the subset of resources to use for data transmission while avoiding resource collision due to other UEs. This article summarizes the resource sensing and selection (RSS) mechanisms for SL RA mode 2 specified in releases 16 and 17 of NR V2X. The critical aspect for RSS is the minimization of UE power consumption during resource sensing, which is caused by multiple blind decodings on the physical SL control channel. To address this issue, the effective number of blind decodings for conventional RSS is quantitatively analyzed to determine the potential enhancements required. Furthermore, an enhanced RA mode 2 procedure based on an intelligent partial-sensing (IPS) scheme is proposed to minimize the number of blind decodings. The proposed IPS utilizes a convolutional neural network-based physical channel-type classification model. Simulation and numerical results show that the throughput obtained with the proposed IPS scheme approximates that obtained via full sensing-based RA mode 2, while reducing the number of blind decodings by 90%.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - ETC > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/90820)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.