A Hybrid Clustering Strategy for Transform Domain Communication System

In recent years, the various Internet of Things (IoT) communication technologies have been researched. The third generation partnership project (3GPP) is dedicated to building the IoT platforms, including NB-IoT, eMTC, and URLLC. However, these platforms cannot adequately meet all the needs of futur...

Full description

Bibliographic Details
Main Authors: Shiyong Ma, Su Hu, Xuezhang Zhu, Qiudi Tang, Qu Luo, Yunfeng Zhou
Format: Article
Language:English
Published: IEEE 2019-01-01
Series:IEEE Access
Subjects:
IoT
Online Access:https://ieeexplore.ieee.org/document/8758430/
Description
Summary:In recent years, the various Internet of Things (IoT) communication technologies have been researched. The third generation partnership project (3GPP) is dedicated to building the IoT platforms, including NB-IoT, eMTC, and URLLC. However, these platforms cannot adequately meet all the needs of future IoT use cases which are usually broad and beyond cellular. In order to reduce the deployment cost, making the full use of the unlicensed spectrum is strongly appealing to the IoT scenarios. Transform domain communication system (TDCS) is a typical system with spectrum sensing and sharing, and it has unique advantages in supporting multiuser communication in a complex electromagnetic environment. Thus, TDCS is proposed for the IoT massive multiple access scenarios. Previous studies indicate that traditional TDCS has poor BER performance when encountering a large number of users, especially with the near-far effect. To address this, a hybrid clustering TDCS targeting at cognitive IoT applications is proposed to support massive multiple access with the near-far effect. The optimal construction method of the sequence set is also given for the hybrid clustering TDCS. Compared with traditional TDCS, the simulation results show that the proposed system achieves an improvement in multiple access ability, and the system performance can be improved by making a compromise on the number of clusters.
ISSN:2169-3536