Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/141303
Title: Deep learning based approach for channel estimation of CP-free OFDM system
Authors: Xiao, Xinhao
Keywords: Engineering::Electrical and electronic engineering::Wireless communication systems
Issue Date: 2020
Publisher: Nanyang Technological University
Abstract: In traditional orthogonal frequency division multiplexing (OFDM) system, the cyclic prefix (CP) is used as a guard interval between two successive symbols to overcome the inter-symbol interference (ISI). In addition, it repeats the end of symbol so that the linear convolution of multipath channel can be modelled recycled. In this dissertation, we aim at OFDM systems, especially the CP-free OFDM system, which without CP insert between the successive symbols at transmission, and using Deep Learning (DL) based approach to address channel estimation problems. The LSTM neural network is established to improve the simulation performance. Indeed, we also investigate their performance with another two popular methods least square (LS) and minimum mean square error (MMSE) channel estimation, and compared with deep learning based approach under different channel models. The simulation results reveal that the Deep Learning (DL) based method obtains lower Bit Error Rates (BERs) when Signal Noise Ratio (SNR) increases. We will also show that, when using the Deep Learning method, the receiver is robust in various situations, such as CP or CP-free system and different pilots number system. DL based approach has better performance than those competitive algorithms in most time.
URI: https://hdl.handle.net/10356/141303
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Theses

Files in This Item:
File Description SizeFormat 
XiaoXinhaoG1901211G_Dissertation_MSC_finalversion.pdf
  Restricted Access
3.9 MBAdobe PDFView/Open

Page view(s)

284
Updated on Jan 29, 2023

Download(s)

13
Updated on Jan 29, 2023

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.