A Machine Learning Method for Prediction of Multipath Channels

doi:10.12142/ZTECOM.201904003

ZTE Communications ›› 2019, Vol. 17 ›› Issue (4): 12-18.DOI: 10.12142/ZTECOM.201904003

• Special Topic • Previous Articles Next Articles

A Machine Learning Method for Prediction of Multipath Channels

Julian AHRENS¹(), Lia AHRENS¹, Hans D. SCHOTTEN^1,²

^1.German Research Center for Artificial Intelligence, Kaiserslautern 67663, Germany
^2.Technical University of Kaiserslautern, Kaiserslautern 67663, Germany

Received:2019-09-30 Online:2019-12-25 Published:2020-04-16
About author:Julian AHRENS(Julian. Ahrens@dfki.de) received his Master’s degreein mathematics from Kiel University (CAU), Germany, while working innon-commutative harmonic analysis. He is currently working as a research?er at the Intelligent Networks Group of Prof. Hans D. Schotten at the Ger?man Research Center for Artificial Intelligence, where he is involved in theBMBF project Future Industrial Network Architecture (FIND). His re?search interests include high performance computing, artificial intelligence,digital signal processing, and harmonic and functional analysis.|Lia AHRENSreceived her Ph. D. degree in stochastics and financialmathematics from Kiel University (CAU), Germany. She is currently work?ing as a senior researcher at the Intelligent Networks Group of Prof. HansD. Schotten at the German Research Center for Artificial Intelligence,where she is involved in the BMBF project TACNET 4. 0―Taktiles Inter?net. Her research interests include stochastic processes, stochastic filter?ing, and machine learning for time series analysis.|Hans D. SCHOTTENreceived his Diploma and Ph. D. degrees in elec?trical engineering from the RWTH Aachen University of Technology, Ger?many. He is a full professor and director of the Institute for Wireless Com?munications and Navigation at the Technical University of Kaiserslautern,Germany. In addition, he is a scientific director of the German ResearchCenter for Artificial Intelligence (DFKI) and head of the department for In?telligent Networks. Since 2018, he is the chairman of the German Societyfor Information Technology and member of the Supervisory Board of theVDE.

Abstract

Abstract:

In this paper, a machine learning method for predicting the evolution of a mobile communication channel based on a specific type of convolutional neural network is developed and evaluated in a simulated multipath transmission scenario. The simulation and channel estimation are designed to replicate real-world scenarios and common measurements supported by reference signals in modern cellular networks. The capability of the predictor meets the requirements that a deployment of the developed method in a radio resource scheduler of a base station poses. Possible applications of the method are discussed.

Key words: channel estimation, channel prediction, convolutional neural network, machine learning, multipath transmission

Julian AHRENS, Lia AHRENS, Hans D. SCHOTTEN. A Machine Learning Method for Prediction of Multipath Channels[J]. ZTE Communications, 2019, 17(4): 12-18.

Figures/Tables 9

Figure 1 Example configuration of the designed simulation.

Figure 2 Example plot of estimated transfer functions.

Table 1 Layout of the 2D CNN for m-step ahead prediction

Layer number	Layer type	Channel sizes	Kernel Size (t, f?)	Dilation parameter	Residual layer

0	Dilated convolution	$2 → 6$	$(4, ? 5)$	$(1, ? 1)$	$1 × 1$ convolution
1	Dilated convolution	$6 → 12$	$(4, ? 5)$	$(4, ? 1)$	$1 × 1$ convolution
2	Dilated convolution	$12 → 12$	$(4, ? 5)$	$(16, ? 1)$	$1 × 1$ convolution
3	Dilated convolution	$12 → 6$	$(4, ? 5)$	$(64, ? 1)$	$1 × 1$ convolution
4	Convolution	$6 → 2 m$	$(1, ? 1)$	$(1, ? 1)$	None

Table 1 Layout of the 2D CNN for m-step ahead prediction

Layer number	Layer type	Channel sizes	Kernel Size (t, f?)	Dilation parameter	Residual layer

0	Dilated convolution	$2 → 6$	$(4, ? 5)$	$(1, ? 1)$	$1 × 1$ convolution
1	Dilated convolution	$6 → 12$	$(4, ? 5)$	$(4, ? 1)$	$1 × 1$ convolution
2	Dilated convolution	$12 → 12$	$(4, ? 5)$	$(16, ? 1)$	$1 × 1$ convolution
3	Dilated convolution	$12 → 6$	$(4, ? 5)$	$(64, ? 1)$	$1 × 1$ convolution
4	Convolution	$6 → 2 m$	$(1, ? 1)$	$(1, ? 1)$	None

Figure 3 Structure of the dilated convolutional layers along the t-axis.

Table 2 Mean squared errors (MSEs) for prediction length ΔtΔt from 1 to 10

$Δ t$	Training	Validation	Test	Trivial

1	0.1466	0.1460	0.1424	0.2217
2	0.1579	0.1569	0.1538	0.2532
3	0.1753	0.1735	0.1707	0.3047
4	0.2014	0.1988	0.1962	0.3750
5	0.2357	0.2321	0.2298	0.4621
6	0.2757	0.2723	0.2704	0.5644
7	0.3249	0.3208	0.3183	0.6788
8	0.3821	0.3763	0.3735	0.8040
9	0.4459	0.4393	0.4354	0.9382
10	0.5133	0.5055	0.5005	1.0797

Table 2 Mean squared errors (MSEs) for prediction length ΔtΔt from 1 to 10

$Δ t$	Training	Validation	Test	Trivial

1	0.1466	0.1460	0.1424	0.2217
2	0.1579	0.1569	0.1538	0.2532
3	0.1753	0.1735	0.1707	0.3047
4	0.2014	0.1988	0.1962	0.3750
5	0.2357	0.2321	0.2298	0.4621
6	0.2757	0.2723	0.2704	0.5644
7	0.3249	0.3208	0.3183	0.6788
8	0.3821	0.3763	0.3735	0.8040
9	0.4459	0.4393	0.4354	0.9382
10	0.5133	0.5055	0.5005	1.0797

Figure 4 Power density spectrum in dB for Δt=1.

Figure 5 Power density spectrum in dB for Δt=2.

Figure 6 Power density spectrum in dB for Δt=5.

Figure 7 Power density spectrum in dB for Δt=10.

References 10

1	BEARD C, STALLINGS W. Wireless Communication Networks and Systems, Global Edition [M]. London, UK: Pearson, 2015
2	COST 207 Management Committee. COST 207 Digital land mobile radio communications [R]. Brussels, Belgium; Luxembourg, Luxembourg: Commission of the European Communities, 1990
3	OORD A VAN DEN, DIELEMAN S, ZEN H, et al. WaveNet: A Generative Model for Raw Audio [EB/OL]. (2016-09-19).
4	SCHULZE H. Stochastische Modelle und digitale Simulation von Mobilfunkkanälen [C]//Kleinheubacher Berichte. Kleinheubach, Germany, 1988: 473–483.
5	K-D KAMMEYER. Nachrichtenübertragung [M]. Wiesbaden, Germany: Vieweg+Teubner Verlag, 2004. DOI:10.1007/978-3-322-94062-9 DOI
6	GOODFELLOW I, BENGIO Y, COURVILLE A. Deep Learning[M]. Cambridge, USA: MIT Press, 2016
7	HOCHREITER S, SCHMIDHUBER J. Long Short-Term Memory [J]. Neural Computation, 1997, 9(8): 1735–1780. DOI: 10.1162/neco.1997.9.8.1735 DOI
8	RUMELHART D E, HINTON G E, WILLIAMS R J. Learning Representations by Back-Propagating Errors [J]. Nature, 1986, 323: 533–536. DOI: 10.1038/323533a0 DOI
9	HE K, ZHANG X, REN S, et al. Deep Residual Learning for Image Recognition [C]//29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR2016. Vegas Las, USA, 2016. DOI:10.1109/CVPR.2016.90 DOI
10	KINGMA D P, BA J L. Adam: A Method for Stochastic Optimization [C]//3rd International Conference on Learning Representations, ICLR 2015. San Diego , USA， 2015. arXiv:1412.6980 [cs.LG]

A Machine Learning Method for Prediction of Multipath Channels

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 9

References 10

Related Articles 0

Recommended Articles

Metrics