link adaptation techniques for future terrestrial and
play

Link Adaptation Techniques for Future Terrestrial and Satellite - PowerPoint PPT Presentation

Link Adaptation Techniques for Future Terrestrial and Satellite Communications Anxo Tato Arias Supervised by Carlos Mosquera Nartallo atlanTTic Research Center, Universidade de Vigo December 13, 2019 Anxo Tato Arias 1 / 71 1. Motivation


  1. 3. Multibeam Satellite Systems with Linear Precoding Simulation Results SINR Absolute Error (Aggregated Results) Global margin [245 beams REAL Null.] 10 0 Comparison maximum SINR error 2.5 Nominal C/N Nominal C/N minus 3 dB 10 -1 2 10 -2 CCDF: P(error > x) Maximum SINR error (dB) 1.5 10 -3 1 10 -4 0.5 10 -5 Nominal C/N Nominal C/N minus 3 dB 10 -6 0 SYNC I/N -10 dB REAL Null. SYNC I/N -20 dB SYNC I/N -25 dB -1.5 -1 -0.5 0 0.5 1 1.5 2 SINR absolute error (dB) Averaging more pilots reduces the nullification threshold and the errors CCDF allows to obtain the margin for a given target FER Anxo Tato Arias 25 / 71

  2. 3. Multibeam Satellite Systems with Linear Precoding Simulation Results SINR Absolute Error at a Fixed Position Error much lower than the maximum in the vast majority of the positions Stationary behavior of SINR and SINR error Anxo Tato Arias 26 / 71

  3. 3. Multibeam Satellite Systems with Linear Precoding Simulation Results System Throughput and Back-off Margin Throughput comparison 100 Perfect CSIT Null. User margin Null. Beam margin Null. Global margin 95 90 Relative Throughput (%) 85 80 75 70 65 60 6 sectors per beam 4 sectors per beam No interbeam scheduling Global margin : 79 % throughput Margin per beam : 84 % throughput Margin per user : 94 % throughput Anxo Tato Arias 27 / 71

  4. 3. Multibeam Satellite Systems with Linear Precoding Simulation Results Nullification Countermeasure: Adaptive Margin per User [1] R. A. Delgado, K. Lau, R. Middleton, R. S. Karlsson, T. Wigren, and Y. Sun. Fast convergence outer loop link adaptation with infrequent updates in steady state . In 2017 IEEE 86th Vehicular Technology Conference (VTC-Fall). Margin evolution 0.2 Simulations 0 Without fading -0.2 Rice fading K = 25 dB -0.4 -0.6 (terrestrial) -0.8 Rice fading K = 34 dB -1 (aeronautical) -1.2 Experimental FER -1.4 80-120 % of the target FER -1.6 0 1 2 3 4 5 Frame number 10 4 Anxo Tato Arias 28 / 71

  5. 3. Multibeam Satellite Systems with Linear Precoding Conclusions Conclusions The practical problem of the nullification in multibeam precoded systems was analyzed A solution based on a link adaptation algorithm was proposed The adaptive margin per user allows to meet the FER constraint with a small impact on the throughput of the system Anxo Tato Arias 29 / 71

  6. 4. Spatial Modulation Transmission Capacity Outline 1. Motivation 1 2. Mobile Satellite System: Field Trial Results 2 3. Multibeam Satellite Systems with Linear Precoding 3 4. Spatial Modulation Transmission Capacity 4 5. Mobile Satellite Systems with Dual Polarization 5 6. Spatial Modulation Systems 6 7. Conclusions 7 Anxo Tato Arias 30 / 71

  7. 4. Spatial Modulation Transmission Capacity Chapter 4 Evaluation of the Spatial Modulation Transmission Capacity Publications Mobile Satellite Systems Fixed Satellite Future terrestrial Systems and satellite communication systems (submitted, under revision) Anxo Tato Arias 31 / 71

  8. 4. Spatial Modulation Transmission Capacity Introduction What is Spatial Modulation? Spatial Multiplexing (SMux) Spatial Multiplexing N t Radio Frequency (RF) chains Max. spectral efficiency: η = N t log 2 M Anxo Tato Arias 32 / 71

  9. 4. Spatial Modulation Transmission Capacity Introduction What is Spatial Modulation? Spatial Multiplexing (SMux) Spatial Modulation (SM) 0 Spatial Spatial Modulation Multiplexing 1 S2=0/1 One RF chain N t Radio Frequency (RF) chains Max. spectral efficiency: Max. spectral efficiency: η = log 2 N t + log 2 M η = N t log 2 M Anxo Tato Arias 32 / 71

  10. 4. Spatial Modulation Transmission Capacity Introduction What is Spatial Modulation? Spatial Multiplexing (SMux) Spatial Modulation (SM) 0 Spatial Spatial Modulation Multiplexing 1 S2=0/1 One RF chain N t Radio Frequency (RF) chains Max. spectral efficiency: Max. spectral efficiency: η = log 2 N t + log 2 M η = N t log 2 M Generalized Spatial Modulation (GSM) Antenna selection R < N t RF chains bits Bit Channel splitter Max. spectral efficiency: M-QAM � N t � η = ⌊ log 2 ⌋ + log 2 M R RF switches Anxo Tato Arias 32 / 71

  11. 4. Spatial Modulation Transmission Capacity Introduction Problem Formulation Maximum Capacity calculation achievable Channel rate Applications: Adaptation of transmission bit rate in adaptive (G)SM systems Theoretical performance evaluation of (G)SM systems Anxo Tato Arias 33 / 71

  12. 4. Spatial Modulation Transmission Capacity Introduction Problem Formulation Maximum Capacity calculation achievable Channel rate Applications: Adaptation of transmission bit rate in adaptive (G)SM systems Theoretical performance evaluation of (G)SM systems Solutions in the literature: Expressions for obtaining the capacity with numerical integration Two analytical approximations of the SM capacity Comparison: Numerical integration: accurate but very time consuming Approximations: reduce notably the time calculation but less accurate Anxo Tato Arias 33 / 71

  13. 4. Spatial Modulation Transmission Capacity Introduction Problem Formulation Maximum Capacity calculation achievable Channel rate Applications: Adaptation of transmission bit rate in adaptive (G)SM systems Theoretical performance evaluation of (G)SM systems Solutions in the literature: Expressions for obtaining the capacity with numerical integration Two analytical approximations of the SM capacity Comparison: Numerical integration: accurate but very time consuming Approximations: reduce notably the time calculation but less accurate Our proposal: Machine Learning (ML) based capacity calculation Anxo Tato Arias 33 / 71

  14. 4. Spatial Modulation Transmission Capacity Introduction System Model SM: y = √ γ · H · x + w = √ γ · h l · s + w h l : Column of the channel matrix H , l = 1 , 2 , . . . , N t Anxo Tato Arias 34 / 71

  15. 4. Spatial Modulation Transmission Capacity Introduction System Model SM: y = √ γ · H · x + w = √ γ · h l · s + w h l : Column of the channel matrix H , l = 1 , 2 , . . . , N t GSM: � � y = γ/ R · H · x + w = γ/ R · H · A l · 1 · s + w A l : Antenna activation pattern matrix from the set A T   1 0 0 0 0 0 A = 0 1 0 0 0 0   0 0 0 0 1 0 Set of L = |A| = 2 ⌊ log 2 ( Nt R ) ⌋ matrices � γ/ R · H · A l · 1 Sum of R columns of H : c l = Anxo Tato Arias 34 / 71

  16. 4. Spatial Modulation Transmission Capacity Introduction Mutual Information and Capacity Expressions Mutual Information (MI) or constrained capacity of SM: � h l s − h l ′ s ′ + w 2   � �   � � √ γ + � w � 2 − γ  � �  1   � � � 2 � 2 I T = log 2 (2 M ) − � log 2 � l ′ =1 e l =1 E W  �  s ∈S s ′ ∈S 2 M       Capacity of GSM: � 1 � C GSM = − 1 � L � L � y CN ( 0 , Φ i ) log 2 j =1 CN ( 0 , Φ j ) d y − log 2 det( π eI N r ) i =1 L L 5 MI or contrained capacity (QPSK) 4.5 MI or contrained capacity (16QAM) Unconstrained capacity (Shannon) 4 3.5 3 (bpcu) 2.5 2 1.5 1 0.5 0 -20 -10 0 10 20 30 SINR (dB) Anxo Tato Arias 35 / 71

  17. 4. Spatial Modulation Transmission Capacity Introduction Analytical Approximations to the SM MI Taylor approximation Jensen approximation Henarejos et al. Guo et al. Drawbacks of these approximations: Biased and limited accuracy Complexity scales with the square of the constellation size M and the number of antennas N t Calculation for a single constellation Anxo Tato Arias 36 / 71

  18. 4. Spatial Modulation Transmission Capacity Neural Network-based MI and Capacity Estimation Proposed Solution Spatial Modulation Neural network MI QPSK Features MI 8PSK Channel extraction MI 16QAM Anxo Tato Arias 37 / 71

  19. 4. Spatial Modulation Transmission Capacity Neural Network-based MI and Capacity Estimation Proposed Solution Spatial Modulation Neural network MI QPSK Features MI 8PSK Channel extraction MI 16QAM Generalized Spatial Modulation Neural network Unconstrained Features Channel capacity extraction Anxo Tato Arias 37 / 71

  20. 4. Spatial Modulation Transmission Capacity Neural Network-based MI and Capacity Estimation Proposed Solution Spatial Modulation Neural network MI QPSK Features MI 8PSK Channel extraction MI 16QAM Generalized Spatial Modulation Neural network Unconstrained Features Channel capacity extraction Anxo Tato Arias 37 / 71

  21. 4. Spatial Modulation Transmission Capacity Neural Network Features Selection Features selection for 2 × 2 SM � � h 1 � 2 D S � D L → D = {� h l s k − h l ′ s k ′ � 2 } = X = { h l · s k } − � h 2 � 2 D S D t L D L can be characterized with four real values: � h 1 � 2 , � h 2 � 2 and h H 1 h 2 � Θ H ∈ [0 , π/ 2] − → Hermitian angle h H 1 h 2 = � h 1 � · � h 2 � · cos Θ H · e i ϕ , ϕ ∈ [ − π, π ] → Kasner’s pseudoangle High SNR Moderate SNR Received symbols "real" SM-BPSK SNR = 10 dB Received symbols "real" SM-BPSK SNR = 15 dB 8 8 6 6 4 4 Non 2 2 Antenna 2 Antenna 2 0 0 orthogonal -2 -2 -4 -4 -6 -6 -8 -8 -8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 Antenna 1 Antenna 1 Received symbols "real" SM-BPSK SNR = 10 dB Received symbols "real" SM-BPSK SNR = 15 dB 8 8 6 6 4 4 2 2 Orthogonal Antenna 2 Antenna 2 0 0 -2 -2 -4 -4 -6 -6 -8 -8 -8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 Antenna 1 Antenna 1 Anxo Tato Arias 38 / 71

  22. 4. Spatial Modulation Transmission Capacity Neural Network Features Selection Generalization for More Antennas and GSM Features for GSM X = { H · A l · 1 · s k } � h l − → c l = γ/ R · H · A l · 1 Number of features # norms # pairs of angles # norms # pairs of angles � N t � SM N t SM 8 × 8 8 28 2 L = 2 ⌊ log 2 ( Nt R ) ⌋ GSM 8 × 8 , R = 2 16 L 2 = 120 � L GSM L 2 = � 2 Features reduction Characterize norms and angles distribution with Q quantiles, equispaced in [0,1] Example Q=5: Minimum, 25th percentile, median, 75th percentile, maximum Anxo Tato Arias 39 / 71

  23. 4. Spatial Modulation Transmission Capacity Simulation Results Simulations setup Supervised learning ML dataset 50 , 000 Rayleigh distributed random matrices h ij ∼ CN (0 , 1) SNR γ ∼ U( − 20 , 20) dB 70 % training, 15 % validation, 15 % testing Calculation reference values of MI and capacity Monte Carlo simulation with 5 , 000 realizations of the noise w Variance in the estimation ∼ 10 − 5 Learning algorithm Levenberg-Marquardt (LM) backpropagation algorithm MSE as cost function Random initialization of weights and margins Architecture One hidden layer of 10 or 20 neurons with sigmoid activation function Linear output Anxo Tato Arias 40 / 71

  24. 4. Spatial Modulation Transmission Capacity Simulation Results Simulation Results: MI of 2 × 2 SM Impact of different input features Option # Features Global MSE 6 . 98 · 10 − 4 i) Column norms and scalar product 4 3 . 36 · 10 − 4 ii) Column norms and angles 4 5 . 21 · 10 − 5 iii) Column norms and distances 6 4 . 96 · 10 − 5 iv) Column norms, distances and scalar product 8 2 . 97 · 10 − 5 v) Column norms, distances and angles 8 Anxo Tato Arias 41 / 71

  25. 4. Spatial Modulation Transmission Capacity Simulation Results Simulation Results: MI of 2 × 2 SM Impact of different input features Option # Features Global MSE 6 . 98 · 10 − 4 i) Column norms and scalar product 4 3 . 36 · 10 − 4 ii) Column norms and angles 4 5 . 21 · 10 − 5 iii) Column norms and distances 6 4 . 96 · 10 − 5 iv) Column norms, distances and scalar product 8 2 . 97 · 10 − 5 v) Column norms, distances and angles 8 Comparison with analytical approximations Anxo Tato Arias 41 / 71

  26. 4. Spatial Modulation Transmission Capacity Simulation Results Simulation Results: MI of 2 × 2 SM Impact of different input features Option # Features Global MSE 6 . 98 · 10 − 4 i) Column norms and scalar product 4 3 . 36 · 10 − 4 ii) Column norms and angles 4 5 . 21 · 10 − 5 iii) Column norms and distances 6 4 . 96 · 10 − 5 iv) Column norms, distances and scalar product 8 2 . 97 · 10 − 5 v) Column norms, distances and angles 8 Comparison with analytical approximations QPSK Global MSE 3 σ Max. error 1 . 87 · 10 − 2 Taylor approximation 0 . 330 0 . 523 1 . 21 · 10 − 2 Jensen based approximation 0 . 229 0 . 300 2 . 97 · 10 − 5 Neural network 0 . 016 0 . 067 Anxo Tato Arias 41 / 71

  27. 4. Spatial Modulation Transmission Capacity Simulation Results Simulation Results: MI of 2 × 2 SM Impact of different input features Option # Features Global MSE 6 . 98 · 10 − 4 i) Column norms and scalar product 4 3 . 36 · 10 − 4 ii) Column norms and angles 4 5 . 21 · 10 − 5 iii) Column norms and distances 6 4 . 96 · 10 − 5 iv) Column norms, distances and scalar product 8 2 . 97 · 10 − 5 v) Column norms, distances and angles 8 Comparison with analytical approximations QPSK Global MSE 3 σ Max. error 1 . 87 · 10 − 2 Taylor approximation 0 . 330 0 . 523 1 . 21 · 10 − 2 Jensen based approximation 0 . 229 0 . 300 2 . 97 · 10 − 5 Neural network 0 . 016 0 . 067 Computational complexity Taylor approx. Jensen approx. Neural network Real products 7 , 168 32 , 800 368 Non linear operations 784 1 , 347 20 Anxo Tato Arias 41 / 71

  28. 4. Spatial Modulation Transmission Capacity Simulation Results Generalization MI of SM with 4 and 8 antennas # features Global MSE 3 . 36 · 10 − 4 SM 2 × 2 option (ii) 4 2 . 40 · 10 − 4 SM 4 × 4 16 5 . 06 · 10 − 5 SM 8 × 8 (Q = 5) 18 Anxo Tato Arias 42 / 71

  29. 4. Spatial Modulation Transmission Capacity Simulation Results Generalization MI of SM with 4 and 8 antennas # features Global MSE 3 . 36 · 10 − 4 SM 2 × 2 option (ii) 4 2 . 40 · 10 − 4 SM 4 × 4 16 5 . 06 · 10 − 5 SM 8 × 8 (Q = 5) 18 Dual Polarization Mobile Satellite Channel: PMod Global MSE 7 . 40 · 10 − 5 Anxo Tato Arias 42 / 71

  30. 4. Spatial Modulation Transmission Capacity Simulation Results Generalization MI of SM with 4 and 8 antennas # features Global MSE 3 . 36 · 10 − 4 SM 2 × 2 option (ii) 4 2 . 40 · 10 − 4 SM 4 × 4 16 5 . 06 · 10 − 5 SM 8 × 8 (Q = 5) 18 Dual Polarization Mobile Satellite Channel: PMod Global MSE 7 . 40 · 10 − 5 Correlated channels Performance degrades with increasing correlation Anxo Tato Arias 42 / 71

  31. 4. Spatial Modulation Transmission Capacity Simulation Results Generalization MI of SM with 4 and 8 antennas # features Global MSE 3 . 36 · 10 − 4 SM 2 × 2 option (ii) 4 2 . 40 · 10 − 4 SM 4 × 4 16 5 . 06 · 10 − 5 SM 8 × 8 (Q = 5) 18 Dual Polarization Mobile Satellite Channel: PMod Global MSE 7 . 40 · 10 − 5 Correlated channels Performance degrades with increasing correlation Capacity of GSM Studied scenarios SM with 2, 4 and 8 antennas GSM with 6 and 8 antennas and 2 or 3 RF chains Results MSE ∼ 10 − 4 3 σ ∼ 0 . 05 Number of neural network inputs: 4 - 27 Anxo Tato Arias 42 / 71

  32. 4. Spatial Modulation Transmission Capacity Conclusions Conclusions A Machine Learning-based solution was proposed for obtaining the capacity of SM and GSM systems Simple neural networks outperform approximations of the literature both in terms of accuracy and complexity The fast and accurate calculation can find applications in adaptive SM systems Anxo Tato Arias 43 / 71

  33. 5. Mobile Satellite Systems with Dual Polarization Outline 1. Motivation 1 2. Mobile Satellite System: Field Trial Results 2 3. Multibeam Satellite Systems with Linear Precoding 3 4. Spatial Modulation Transmission Capacity 4 5. Mobile Satellite Systems with Dual Polarization 5 6. Spatial Modulation Systems 6 7. Conclusions 7 Anxo Tato Arias 44 / 71

  34. 5. Mobile Satellite Systems with Dual Polarization Chapter 5 Link Adaptation in Mobile Satellite Systems with Dual Polarization Publications Mobile Satellite Systems Fixed Satellite Future terrestrial Systems and satellite communication systems Anxo Tato Arias 45 / 71

  35. 5. Mobile Satellite Systems with Dual Polarization Motivation Introduction User link Feeder link RHCP f (MHz) 1518 1626.5 1675 1559 Mobile Terminal (MT) Satcom Transceiver Gateway Anxo Tato Arias 46 / 71

  36. 5. Mobile Satellite Systems with Dual Polarization Motivation Introduction User link Feeder link RHCP f (MHz) 1518 1626.5 1675 1559 LHCP Mobile Terminal (MT) DP Satcom Transceiver Gateway Link adaptation in... Mobile Satellite Communications with Dual Polarization (DP) L-band (1-2 GHz) and S-band (2-4 GHz) RHCP and LHCP as a 2 × 2 MIMO Several MIMO modes and MCS available Anxo Tato Arias 46 / 71

  37. 5. Mobile Satellite Systems with Dual Polarization System Model Four Available MIMO Modes SISO Anxo Tato Arias 47 / 71

  38. 5. Mobile Satellite Systems with Dual Polarization System Model Four Available MIMO Modes SISO Orthogonal Polarization-Time Block Code (OPTBC) ∼ Alamouti Polarization-time L codeword QPSK info bits Variable rate OPTBC channel encoder (Alamouti) r R 2 symbol periods Anxo Tato Arias 47 / 71

  39. 5. Mobile Satellite Systems with Dual Polarization System Model Four Available MIMO Modes SISO Orthogonal Polarization-Time Block Code (OPTBC) ∼ Alamouti Polarization-time L codeword QPSK info bits Variable rate OPTBC channel encoder (Alamouti) r R 2 symbol periods Polarized Modulation (PMod) ∼ Spatial Modulation (SM) L QPSK Variable rate channel encoder R info bits Bit splitter Polarization Variable rate mapper channel encoder Anxo Tato Arias 47 / 71

  40. 5. Mobile Satellite Systems with Dual Polarization System Model Four Available MIMO Modes SISO Orthogonal Polarization-Time Block Code (OPTBC) ∼ Alamouti Polarization-time L codeword QPSK info bits Variable rate OPTBC channel encoder (Alamouti) r R 2 symbol periods Polarized Modulation (PMod) ∼ Spatial Modulation (SM) L QPSK Variable rate channel encoder R info bits Bit splitter Polarization Variable rate mapper channel encoder Vertical-Bell Lab. Layered Space-Time (V-BLAST) ∼ Spatial Multiplexing QPSK L Variable rate Serial info bits channel encoder to Parallel r QPSK R Anxo Tato Arias 47 / 71

  41. 5. Mobile Satellite Systems with Dual Polarization System Model Modulation and Coding Schemes (MCS) QPSK constellation 9 coding rates for symbols bits (SISO, OPTBC, PMod, V-BLAST) 0.9 0.8 MCS Coding rate 0.7 0.6 0.5 0.4 0.3 -4 -2 0 2 4 6 Effective SNR (dB) 9 coding rates for polarization bits of PMod Coding rate 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Anxo Tato Arias 48 / 71

  42. 5. Mobile Satellite Systems with Dual Polarization System Model Channel Generation M. Sellathurai et al., “Space-time coding in mobile satellite communications using dual-polarized channels”, IEEE Transactions on Vehicular Technology, Jan 2006 H = β e j φ K L + ξ e j φ K S + DK D Anxo Tato Arias 49 / 71

  43. 5. Mobile Satellite Systems with Dual Polarization System Model Physical Layer Abstraction 1 Channel generator: { H n , n = 1 , 2 , . . . , N } 10 5 0 Channel matrix coefficients (dB) -5 -10 -15 -20 -25 h 11 -30 h 22 h 12 -35 h 21 -40 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Time (s) 2 SINR calculation per received symbols Different equation for each MIMO mode 3 SINR compression 13 12 Channel matrix coefficients (dB) 11 10 9 8 7 h 11 6 h 22 h 12 5 h 21 4 500 1000 1500 2000 2500 Anxo Tato Arias 50 / 71

  44. 5. Mobile Satellite Systems with Dual Polarization System Model Link Adaptation Algorithms MIMO mode selection Mode which maximizes spectral efficiency By the receiver MCS selection Using LUT with adaptive margins By the transmitter PMod : two LUTs (independent coding rates for symbols and polarization) Margin MCS adaptation frame i Inner loop Outer loop SISO, OPTBC, V-BLAST. PMod (symbols bits) Anxo Tato Arias 51 / 71

  45. 5. Mobile Satellite Systems with Dual Polarization System Model Link Adaptation Algorithms MIMO mode selection Mode which maximizes spectral efficiency By the receiver MCS selection Using LUT with adaptive margins By the transmitter PMod : two LUTs (independent coding rates for symbols and polarization) 0.9 0.8 Margin Margin Rate MCS adaptation adaptation frame i frame i 0.2 0.1 Inner loop Inner loop Outer loop Outer loop SISO, OPTBC, V-BLAST. PMod (symbols bits) PMod (polarization bits) Anxo Tato Arias 51 / 71

  46. 5. Mobile Satellite Systems with Dual Polarization Simulation Results MIMO Mode Selected mode 100 OPTBC PMod V-BLAST 80 Frequency (%) 60 40 20 0 -5 0 5 10 15 20 25 SNR (dB) Very low SNRs: OPTBC ( ∼ Transmit diversity) Low SNRs: PMod Moderate and high SNRs: V-BLAST ( ∼ Spatial Multiplexing) Anxo Tato Arias 52 / 71

  47. 5. Mobile Satellite Systems with Dual Polarization Simulation Results Spectral Efficiency 3.5 SISO 3 Spectral Efficiency (bps/Hz) 2.5 2 1.5 1 0.5 0 -5 0 5 10 15 20 25 SNR (dB) Anxo Tato Arias 53 / 71

  48. 5. Mobile Satellite Systems with Dual Polarization Simulation Results Spectral Efficiency 3.5 SISO OPTBC 3 Spectral Efficiency (bps/Hz) 2.5 2 1.5 1 0.5 0 -5 0 5 10 15 20 25 SNR (dB) • Operation at lower SNRs and better spectral efficiency Anxo Tato Arias 53 / 71

  49. 5. Mobile Satellite Systems with Dual Polarization Simulation Results Spectral Efficiency 3.5 SISO OPTBC 3 PMod Spectral Efficiency (bps/Hz) 2.5 2 1.5 1 0.5 0 -5 0 5 10 15 20 25 SNR (dB) • + 50 % spectral efficiency Anxo Tato Arias 53 / 71

  50. 5. Mobile Satellite Systems with Dual Polarization Simulation Results Spectral Efficiency 3.5 SISO OPTBC 3 PMod Spectral Efficiency (bps/Hz) V-BLAST 2.5 2 1.5 1 0.5 0 -5 0 5 10 15 20 25 SNR (dB) • + 100 % spectral efficiency Anxo Tato Arias 53 / 71

  51. 5. Mobile Satellite Systems with Dual Polarization Simulation Results Spectral Efficiency and Frame Error Rate 3.5 10 0 SISO SISO OPTBC + V-BLAST OPTBC + V-BLAST 3 OPTBC + V-BLAST + PMod OPTBC + V-BLAST + PMod 10 -1 Spectral Efficiency (bit/s/Hz) 2.5 10 -2 2 FER 1.5 10 -3 1 10 -4 0.5 0 10 -5 -5 0 5 10 15 20 25 -5 0 5 10 15 20 25 SNR (dB) SNR (dB) Inclusion of PMod improves efficiency at low SNRs Target FER is satisfied Anxo Tato Arias 54 / 71

  52. 5. Mobile Satellite Systems with Dual Polarization Conclusions Conclusions Higher rates can be achieved by exploiting both polarizations MIMO mode and MCS can be adjusted Polarized Modulation increases spectral efficiency at low SNRs Anxo Tato Arias 55 / 71

  53. 6. Spatial Modulation Systems Outline 1. Motivation 1 2. Mobile Satellite System: Field Trial Results 2 3. Multibeam Satellite Systems with Linear Precoding 3 4. Spatial Modulation Transmission Capacity 4 5. Mobile Satellite Systems with Dual Polarization 5 6. Spatial Modulation Systems 6 7. Conclusions 7 Anxo Tato Arias 56 / 71

  54. 6. Spatial Modulation Systems Chapter 6 Deep Learning Assisted Rate Adaptation in Spatial Modulation Links Mobile Satellite Systems Publications Fixed Satellite Future terrestrial Systems and satellite communication systems Anxo Tato Arias 57 / 71

  55. 6. Spatial Modulation Systems Introduction Introduction Spatial Modulation New modulation scheme for 5G and beyond 5G Multi-antenna: high spectral efficiency Low complexity: single RF chain Better energy efficiency Anxo Tato Arias 58 / 71

  56. 6. Spatial Modulation Systems Introduction Introduction Spatial Modulation New modulation scheme for 5G and beyond 5G Multi-antenna: high spectral efficiency Low complexity: single RF chain Better energy efficiency Coding rate selection mechanism for adaptive SM systems Supervised learning Deep neural network Anxo Tato Arias 58 / 71

  57. 6. Spatial Modulation Systems Introduction Introduction Spatial Modulation New modulation scheme for 5G and beyond 5G Multi-antenna: high spectral efficiency Low complexity: single RF chain Better energy efficiency Coding rate selection mechanism for adaptive SM systems Supervised learning Deep neural network SM rate adaptation problem maximize r log 2 ( N t M ) r (4) subject to r ∈ { r 1 , r 2 , . . . , r K } BER( γ ; r , H ) ≤ p 0 . Anxo Tato Arias 58 / 71

  58. 6. Spatial Modulation Systems Introduction Block Diagram of an Adaptive SM System Adaptive SM Transmitter Antenna selection Variable rate Bit splitter channel encoder Information bits M-QAM modulator selected coding rate Feedback channel Neural Network SM Receiver aided coding rate selection Channel coding rate estimation in use Soft LLRs Channel detection decoding Information bits Figure 1: Block diagram of an adaptive SM system with variable coding rate. Anxo Tato Arias 59 / 71

  59. 6. Spatial Modulation Systems Proposed Method DL-based Coding Rate Selection 1 Evaluation of the performance of the channel codes System level simulations BER( γ ; r , H ) Figure 2: The different channel codes performance must be evaluated for a large number of channel matrices. Anxo Tato Arias 60 / 71

  60. 6. Spatial Modulation Systems Proposed Method DL-based Coding Rate Selection 2 Extraction of the SNR thresholds 3 2.5 Spectral efficiency (bits/s/Hz) 2 1.5 1 0.5 -5 0 5 10 15 Required SNR (dB) Figure 3: The minimum required SNR to guarantee a given BER p 0 with each coding rate for a set of 20 different channel matrices. Anxo Tato Arias 61 / 71

  61. 6. Spatial Modulation Systems Proposed Method DL-based Coding Rate Selection 3 Building the dataset for Machine Learning Dataset X = { ( x i , y i ) , i = 1 , 2 , . . . , m } Neural network input features: � t γ � h 1 � 2 , γ � h 2 � 2 � � � x = g ( γ, H ) = sort , Θ H , ϕ Neural network output variable: y = r k (target coding rate) Anxo Tato Arias 62 / 71

  62. 6. Spatial Modulation Systems Proposed Method DL-based Coding Rate Selection 3 Building the dataset for Machine Learning Dataset X = { ( x i , y i ) , i = 1 , 2 , . . . , m } Neural network input features: � t γ � h 1 � 2 , γ � h 2 � 2 � � � x = g ( γ, H ) = sort , Θ H , ϕ Neural network output variable: y = r k (target coding rate) 4 Neural network training Training (70 %) and validation (15 %) datasets Anxo Tato Arias 62 / 71

  63. 6. Spatial Modulation Systems Proposed Method DL-based Coding Rate Selection 3 Building the dataset for Machine Learning Dataset X = { ( x i , y i ) , i = 1 , 2 , . . . , m } Neural network input features: � t γ � h 1 � 2 , γ � h 2 � 2 � � � x = g ( γ, H ) = sort , Θ H , ϕ Neural network output variable: y = r k (target coding rate) 4 Neural network training Training (70 %) and validation (15 %) datasets 5 Performance evaluation Testing dataset (15 %) Confussion matrix: accuracy, rate of under-selection, outage probability Anxo Tato Arias 62 / 71

  64. 6. Spatial Modulation Systems Proposed Method DL-based Coding Rate Selection 3 Building the dataset for Machine Learning Dataset X = { ( x i , y i ) , i = 1 , 2 , . . . , m } Neural network input features: � t γ � h 1 � 2 , γ � h 2 � 2 � � � x = g ( γ, H ) = sort , Θ H , ϕ Neural network output variable: y = r k (target coding rate) 4 Neural network training Training (70 %) and validation (15 %) datasets 5 Performance evaluation Testing dataset (15 %) Confussion matrix: accuracy, rate of under-selection, outage probability 6 Operation phase Coding rate selection with fixed neural network parameters θ Anxo Tato Arias 62 / 71

  65. 6. Spatial Modulation Systems Simulation Results Simulated System Parameters SM 2 × 2 with QPSK constellation and 9 coding rate options Paramter Value Transmit and receive antennas N t = 2, N r = 2 Constellation QPSK ( M = 4) Channel coding DVB-S2 codes (BCH + LDPC) Coding rate options 1/4, 1/3, 2/5, 1/2, 3/5, 2/3, 3/4, 5/6, 9/10 p 0 = 10 − 4 Target BER Channel matrices 1000 Rayleigh ditributed SNR range − 5 to 15 dB (0 . 5 dB steps) Neural network configuration • Three hidden layers: 20+15+10 neurons • Activation function: tangent hyperbolic • Output layer: linear Anxo Tato Arias 63 / 71

  66. 6. Spatial Modulation Systems Simulation Results Classification Performance 1 10 Y=X Y=X Points Points 0.8 8 Calculated coding rate index Calculated coding rate 0.6 6 0.4 4 0.2 2 0 -0.2 0 0 0.2 0.4 0.6 0.8 1 0 2 4 6 8 10 Target coding rate Target coding rate index Margin ∆ = 0 ∆ = 0 . 03 Mean accuracy 96 . 2 % 80 . 0 % Outage 2 . 0 % 0 % Underselection 1 . 7 % 19 . 8 % Anxo Tato Arias 64 / 71

  67. 6. Spatial Modulation Systems Simulation Results SM Link Adaptation Fixed coding rate r = r k , for some fixed k . (5) MI-based coding rate selection � I T − ∆ � � � I T − ∆ � � r = Q = arg min − r k � , (6) � � 3 3 r k � DL-based coding rate selection y − ∆) = arg min r k | ˆ y − ∆ − r k | . r = Q (ˆ (7) Anxo Tato Arias 65 / 71

  68. 6. Spatial Modulation Systems Simulation Results Spectral Efficiency 3 Genie-aided Fixed rate 1/4 Fixed rate 1/2 2.5 Spectral efficiency (bits/s/Hz) 2 1.5 1 0.5 0 -5 0 5 10 15 SNR (dB) Anxo Tato Arias 66 / 71

  69. 6. Spatial Modulation Systems Simulation Results Spectral Efficiency 3 Genie-aided MI-based =0.80 Fixed rate 1/4 2.5 Fixed rate 1/2 Spectral efficiency (bits/s/Hz) 2 1.5 1 0.5 0 -5 0 5 10 15 SNR (dB) Anxo Tato Arias 66 / 71

  70. 6. Spatial Modulation Systems Simulation Results Spectral Efficiency 3 Genie-aided Deep Learning based =0.03 MI-based =0.80 2.5 Fixed rate 1/4 Fixed rate 1/2 Spectral efficiency (bits/s/Hz) 2 1.5 1 0.5 0 -5 0 5 10 15 SNR (dB) Anxo Tato Arias 66 / 71

  71. 6. Spatial Modulation Systems Simulation Results Spectral Efficiency and FER 10 0 3 Genie-aided Deep Learning based =0.03 MI-based =0.80 Deep Learning based =0.03 MI-based =0.80 Fixed rate 1/4 2.5 Fixed rate 1/2 Fixed rate 1/4 Fixed rate 1/2 Spectral efficiency (bits/s/Hz) 10 -1 2 Outage probability 1.5 10 -2 1 0.5 10 -3 0 -5 0 5 10 15 -5 0 5 10 15 SNR (dB) SNR (dB) Anxo Tato Arias 67 / 71

  72. 6. Spatial Modulation Systems Conclusions Conclusions and Future Work Conclusions Neural networks can be used to select the coding rate in adaptive SM systems The spectral efficiency is very close to the maximum achievable value Future work Extension to more general scenarios: more antennas and several constellations Online adaptation of the neural network during the operation Anxo Tato Arias 68 / 71

  73. 7. Conclusions Outline 1. Motivation 1 2. Mobile Satellite System: Field Trial Results 2 3. Multibeam Satellite Systems with Linear Precoding 3 4. Spatial Modulation Transmission Capacity 4 5. Mobile Satellite Systems with Dual Polarization 5 6. Spatial Modulation Systems 6 7. Conclusions 7 Anxo Tato Arias 69 / 71

  74. 7. Conclusions Conclusions Future terrestrial and satellite communication systems Anxo Tato Arias 70 / 71

  75. 7. Conclusions Conclusions Future terrestrial and satellite communication systems Anxo Tato Arias 70 / 71

Recommend


More recommend