.png){kind=icon}
返回
Data-Driven Deep Learning for OTFS Detection br
.png){kind=icon}
PRE
AI
DOI：10.23919/JCC.2023.01.008.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Recently, orthogonal time frequency space(OTFS) was presented to alleviate severe Doppler ef-fects in high mobility scenarios. Most of the cur-rent OTFS detection schemes rely on perfect channelstate information (CSI). However, in real-life systems,the parameters of channels will constantly change,which are often difficult to capture and describe. Inthis paper, we summarize the existing research onOTFS detection based on data-driven deep learning(DL) and propose three new network structures. Thepresented three networks include a residual network(ResNet), a dense network (DenseNet), and a resid-ual dense network (RDN) for OTFS detection. Thedetection schemes based on data-driven paradigms donot require a model that is easy to handle mathemat-ically. Meanwhile, compared with the existing fullyconnected-deep neural network (FC-DNN) and stan-dard convolutional neural network (CNN), these threenew networks can alleviate the problems of gradientexplosion and gradient disappearance. Through simu-lation, it is proved that RDN has the best performanceamong the three proposed schemes due to the combi-nation of shallow and deep features. RDN can solvethe issue of performance loss caused by the traditionalnetwork not fully utilizing all the hierarchical information
Keyword:
data-driven
deep learning
OTFS
detection
