MINISTRY OF EDUCATION AND TRAINING HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY NGUYEN HONG ANH LOCATION-AWARE MULTIPATH-BASED CHANNEL PREDICTION FOR NEXT GENERATION WIRELESS COMMUNICATION SYSTEMS DOCTORAL DISSERTATION OF TELECOMMUNICATIONS ENGINEERING Hanoi−2022 MINISTRY OF EDUCATION AND TRAINING HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY NGUYEN HONG ANH LOCATION-AWARE MULTIPATH-BASED CHANNEL PREDICTION FOR NEXT GENERATION WIRELESS COMMUNICATION SYSTEMS Major: Telecommunication Engineering Code: 9520208 DOCTORAL DISSERTATION OF TELECOMMUNICATIONS ENGINEERING SUPERVISORS: 1. Nguyen Van Khang 2. Klaus Witrisal Hanoi−2022 DECLARATION OF AUTHORSHIP I declare that I have authored this thesis independently, that I have not used other than the declared sources/resources, and that I have explicitly indicated all material which have been quoted either literally or by content from the sources used. Hanoi, / / 2022 PhD Student Nguyen Hong Anh SUPERVISORS Assoc.
Nguyen Van Khang i ACKNOWLEDGEMENT This dissertation was written during my doctoral course at School of Electronics and Telecommunications (SET), Hanoi University of Science and Technology (HUST). I was also received tremendous supports from the Signal Processing and Speech Com-munication Laboratory (SPSC), Graz University of Technology (TUGraz), Austria. I am so grateful for all people who always support and encourage me for completing this study. First, I would like to express my sincere gratitude to my advisors for their effective guidance, their patience, continuous support and encouragement, and their immense knowledge.
I would like to thank all members of SPSC, TUGraz. They have been very kind and supportive during my visits to Graz. They helped me a lot with their deep understand-ing of the group’s topics and researches. I also would like to thank all my colleagues in SET, HUST.
They have always helped me with the research process and given helpful advice for me to overcome my own difficulties. During my Ph.D course, I have received many supports from the Management Board of School of Electronics and Telecommunications. Thanks to my employer, HUST for all necessary support and encouragement during my Ph. I am also grateful to Vietnam’s Program 911, for their generous financial support.
Last but not least, I would like to thank OeAD and SPSC for giving funds for my research visits to Graz. Special thanks to my family and relatives for their never-ending support and sacri-fice. Student ii CONTENTS DECLARATION OF AUTHORSHIP. ix LIST OF TABLES.
xiii LIST OF FIGURES. INTRODUCTION AND MOTIVATION. Location-awareness in mmWave beamforming. Location-awareness in vehicular communications.
Location-awareness in adaptive mobile communications, scheduling and routing. Channel quality metric (CQM). Challenges and motivations. Purposes and objectives.
Towards a site-specific radio propagation modeling. Towards a large-scale predicting of radio channel statistics. Towards a side information-aided single-anchor multipath-based localization 7 1. Contributions and outline.
SIGNAL AND SYSTEM MODELS. Representation of reflectors using virtual anchors (VAs). Floor plan/environment information for location-aware applications. Hybrid geometric/stochastic signal model.
Channel quality indicators. Signal-to-interference-plus-noise ratio (SINR). Position error bound (PEB). Contribution of individual SMCs in the overall channel capacity.
GAUSSIAN PROCESS REGRESSION FOR SMC AMPLITUDES 28 3. SMC propagation model. GP Modeling (GPM) of the SMC Amplitudes. Evaluate the quality of prediction.
Experiment and result. Measurement pre-processing. GPR of SMC Amplitudes. GPR of SMC Phases.
RADIO ENVIRONMENT MAP FOR SITE-SPECIFIC PROPAGATION MODELING. Radio environment map (REM) using Gaussian Process regression (GPR) 47 4. Position error bound. APPLICATION OF GPR - ENABLED REMS TO ROBUST POSITIONING.
Description of channel measurement campaigns. Abbreviation Meaning 1 ACF AutoCorrelation Function 2 ADC Analog-to-Digital Converter 3 AOA Angle-Of-Arrival 4 AOD Angle-Of-Departure 5 AWGNAdditive White Gaussian Noise 6 BER Bit Error Rate 7 BF Beam Forming 8 BS Base Station 9 CDF Cumulative Distribution Function 10 CIR Channel Impulse Response 11 CRLB Cramer Rao Lower Bound 12 CQM Channel Quality Metric 13 CSI Channel State Information 14 DM Diffuse Multipath 15 DMC Diffuse Multipath Component 16 EC Energy Capture 17 ECC European Communications Committee 18 EFIM Equivalent Fisher Information Matrix 19 EPB East Plaster Board 20 FCC Federal Communications Commission 21 FIM Fisher Information Matrix 22 GNSS Global Navigation Satellite System 23 GP Gaussian Process 24 GPM Gaussian Process Model 25 GPR Gaussian Process Regression 26 GPS Global Positioning System 27 GSCM Geometry-based Stochastic Channel Model 28 IoT Internet-of-Thing 29 LIDAR Light Detection And Ranging 30 LLHF Log LikeliHood Function vi 31 M2M Machine-to-Machine 32 MAC Media Access Control 33 MIMO Massive Input Massive Output 34 MINT Multipath-assisted Indoor Navigation and Tracking 35 ML Maximum Likelihood 36 MMSE Minimum Mean Square Error 37 MPC MultiPath Component 38 MRC Maximal Ratio Combining 39 MSLL Mean Square Log Loss 40 NLOS Non-Line-Of-Sight 41 NGW North Glass Window 42 LOS Line-Of-Sight 43 OFDM Orthogonal Frequency Division Multiplexing 44 PA Physical Anchor 45 PAM Pulse Amplitude Modulation 46 PDF Probability Distribution Function 47 PDP Power Delay Profile 48 PHY PHYsical Layer Protocol 49 PEB Position Error Bound 50 QAM Quadrature Amplitude Modulation 51 REM Radio Environment Map 52 RF Radio Frequency 53 RFID Radio Frequency IDentification 54 RRC Root Raised Cosine 55 RSS Received Signal Strength 56 RX Receiver 57 RV Random Variable 58 SALMA Single-Anchor Localization system using Multipath Assistance 59 SEP Symbol Error Probability 60 SIMO Single-Input-Multiple-Output 61 SINR Signal-to-Interference-plus-Noise Ratio 62 SLAM Simultaneous Localization And Mapping 63 SMC Specular Multipath Component 64 SMSE Standard Mean Square Error 65 SNR Signal-to-Noise Ratio vii 66 SW South Wall 67 ToF Time-of-Flight 68 TX Transmitter 69 UE User Equipment 70 URLLC Ultra-Reliable Low-Latency Communication 71 UWB Ultra Wide Band 72 VA Virtual Anchor 73 WW West Wall viii SYMBOLS No. Symbol Meaning 1 a Position of the PA 1 2 a Position of the k-th VA k 3 a GP correlation angle 4 c Speed of light 5 C C C Capacitíe of the whole channel, of individual link and of all the 0, k, all SMC links together 6 c GP covariance GP 7 C Covariance of noise vector 8 ˜ Covariance of the noise vector takes into account the effect of DM C 9 dk Distance to the k-th VA 10 d0 Reference distance 11 d Transmitted symbol 12 D Database 13 ∗ Database for test points D 14 E Energy of the transmitted symbol s 15 f Carrier frequency c 16 GP Gaussian Process model 17 h(t) Channel impulse response 18 ˜ Vector sampled from the filtered impulse response h 19 hk Average channel gain after projection on to the pulse shifted by τk 20 h Vector containing channel gains from projecting onto deterministic shifted pulses 21 i sample time index 22 I Identity matrix 23 Jψ Fisher Information Matrix 24 J Equivalent Fisher Information Matrix p 25 Jr Ranging direction matrix 26 Kν (t) Autocorrelation function of ν(t) 27 k VA or SMC index 28 K Total number of SMCs ix 29 K,k Covariance matrix, vector 30 L Likelihood function 31 N0 Noise power 32 p,p′ Agent position 33 P PEB 34 p∗ Position of test point 35 pˆ Estimated location 36 r(t; p) CIR at agent position p after filtering at the receiver 37 r Filtered CIR in vector notation 38 S (τ) PDP ν 39 s(t) Transmitted UWB pulse 40 ˆ Estimated PDP Sν(.) 41 SINRk SINR of the k-th MPCs 42 s(τk(p)) Shifted pulse in vector form 43 Ts Pulse duration 44 T Effective pulse duration p 45 tr Trace of a matrix 46 w(t) Noise signal 47 w Noise component after projection on to the pulse shifted by τ k k 48 x Abscissa of SMC amplitude 49 x Vector of abscissas of SMC amplitudes 50 x˜ Vector of the residual of abscissas of SMC amplitudes accounting for the SMC variance and the uncertainty of phases 51 a , b Real and imaginary parts of ejφk k k 52 z Decision variable 53 α Amplitude of the k-th MPC k 54 αˆ Estimated SMC complex amplitude 55 β Effective bandwidth of the energy-normalized transmit pulse 56 βabs , βph GP mean for the absolute value and phase of the k-th SMC ampli- k k tude 57 δ(t) Dirac delta function 58 δ ,δ(∥ p − p′ ∥)Kronecker delta function i,j 59 ϵ Measurement noise 60 Γ(.) The combined angle-dependence of the antenna pattern and the reflection coefficient, a. the normalized SMC amplitude 62 λ Langrange multiplier k 63 Λ SMC amplitude’s variance diagonal matrix 64 µ GP mean GP 65 µ Vector of SMC amplitude expected mean obtained via.
GPR 66 νˆ Estimated DM 67 ν DM after projection on to the pulse shifted by τk k 68 ν(t) Impulse response of DM 69 Ω Set of measured agent positions 70 ϕ Direction angle w.t the k-th VA k 71 Π ⊥ Orthogonal complement projection on the subspace spanned by the S (τ) column of S(τ) 72 ψ Abscissa of the normalized MPC amplitude 73 Φ SMC amplitude’s phase diagonal matrix 74 σ Standard deviation of the abscissa of the normalized MPC ampli- tude 75 σ ph Standard deviation of the phase of the normalized MPC amplitude ν 76 σ Standard deviation of the GP correlation kernel 77 σ Standard deviation of noise n 78 σ Standard deviation accounting for DMC ν 79 σ Standard deviation of measurement noise ϵ 80 τk Time arrival/delay for the k-th MPC 81 τ Delay 82 τˆ Estimated delay w.t the k-th VA k 83 θ GP hyper-parameters 84 ˆ Estimated GP hyper-parameters θ 85 φk Phase of the k-th SMC amplitude 86 ζ Phase of the normalized MPC amplitude 87 E{.)∗ Complex conjugate operator 92 ∗ Convolution operator xi 93 ∥. Inner product 95 * of the abscissa of the normalized MPC amplitude ∗abs 96 * of the phase of the normalized MPC amplitude ∗ph 97 ∗H Hermitian transpose xii LIST OF TABLES 2.1 EC, Goodness-of-Fit, SINR for example MPCs .2 Parameters of the GP model for the absolute value of the SMC amplitudes.4 Error percentage of the power prediction of SMCs. 40 xiii LIST OF FIGURES 2.1 Illustration of multipath geometry using VAs for transmission between a PA and a mobile agent with exact floor plan information, as seen in [50, Fig.2 Illustration of the VAs for the PA and an agent with PDF p(VA) and p(agent), respectively, as seen in [51, Figure 1. The VA false represents a false detection of VA .3 Floor plan of the evaluation scenario .4 EC of individual SMC.
number of visible VAs .6 CDF of channel capacity of different channel components .1 GP regression for SMC amplitude .2 GP prediction for SMC amplitude .3 Estimated SMC amplitude compared to measurement and GP statistics .4 GP regression for SMC phases .5 GP prediction for SMC phases .1 GPR on SMC amplitudes .1 GPR on SMC amplitudes (cont.1 GPR on SMC amplitudes (cont.2 Predicted SMC amplitudes .2 Predicted SMC amplitudes (cont.3 GP prediction of SINR for the whole FP .3 GP prediction of SINR for the whole FP (cont.3 GP prediction of SINR for the whole FP (cont.4 GP predicted PEB .1 CDF of localization errors .2 LLHF values at evaluation points .3 CDF of localization errors .1 MINT measurement scenario .2 Photo of corridor scenario .3 Floor plan of the evaluation scenario .4 Equipment used for measurements .5 Calibration setup for time domain measurements. 87 xiv CHAPTER 1 INTRODUCTION AND MOTIVATION 1. Literature review 5G and beyond-5G will be characterized by a wide variety of use cases, as well as orders-of-magnitude increases in mobile data volume per area, number of connected devices, and typical user data rate, all compared to current mobile communication systems [1]. It is visioned that context information in general and location information in particular can complement both traditional and disruptive technologies in addressing several of the challenges in 5G and beyond-5G networks.
A majority of 5G devices will be able to rely on ubiquitous location awareness, supported through several technological developments: a multitude of global naviga- tion satellite systems (GNSS) are being rolled out, complementing the current global positioning system (GPS).
