Design of Adaptive 4D-8PSK-TCM based on CNN for LEO Satellite Communication

Alternative Title
저궤도 위성 통신을 위한 합성곱 신경망 기반의 4D-8PSK-TCM 설계
Author(s)
김송
Alternative Author(s)
Song Kim
Advisor
김재현
Department
일반대학원 AI융합네트워크학과
Publisher
The Graduate School, Ajou University
Publication Year
2021-02
Language
eng
Keyword
4D-8PSK-TCMCNNLEO Satellite Communication
Alternative Abstract
The consultative committee for space data systems (CCSDS) recommends four dimensional eight-ary phase shift keying trellis coded modulation (4D-8PSK-TCM) for a high-speed transmission technology in the X-band. The 4D-8PSK-TCM can reduce computational complexity with T-algorithm that gains benefits by removing elements below the threshold.This thesis designs the digital communication system using 4D-8PSK-TCM and proposes adaptive decoder to maintain the bit error rate (BER) performance and minimize computational complexity for all intervals of energy per bit to noise spectral density ratio (Eb/N0). The designed system is applied with zero termination to prevent degradation of BER performance. The proposed adaptive decoder is applied with Eb/N0 classification model that optimize T-algorithm value by estimation of Eb/N0 of received signal using image generator and convolutional neural network (CNN). Moreover, this thesis compares the conventional 4D-8PSK-TCM and the 4D-8PSK-TCM using proposed adaptive decoder in terms of computational complexity and BER performance. The performance evaluation results show that proposed adaptive decoder can minimize to computational complexity while maintaining BER performance for all intervals of Eb/N0. It is expected that the proposed adaptive decoder can be used in low Earth orbit (LEO) satellite for low latency decoding.
URI
https://dspace.ajou.ac.kr/handle/2018.oak/20012
Fulltext

Appears in Collections:
Graduate School of Ajou University > Department of Artificial Intelligence Convergence Network > 3. Theses(Master)
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

Browse