Time Domain Synchronous OFDM Based on Simultaneous MultiChannel

Time Domain Synchronous OFDM Based on Simultaneous Multi-Channel Reconstruction Linglong Dai, Jintao Wang, Zhaocheng Wang Paschalis Tsiaflakis, Marc Moonen Tsinghua University & KU Leuven 2013 -06 -11

Contents 1 Technical Background 2 Proposed Solution 3 Performance Analysis 4 Simulation Results 5 Conclusions 2

OFDM Transmission Technologies l Cyclic Prefix OFDM (CP-OFDM) and Zero Padding OFDM (ZPOFDM) l Time Domain Synchronous OFDM (TDS-OFDM) – High spectral efficiency (increased by about 10% due to no pilot) – Fast and reliable synchronization CP/ZP-OFDM Symbol CP/ZP-OFDM: CP/ZP Data + Pilots TDS-OFDM Symbol TDS-OFDM: PN Data (No Pilot) (a) Comparison in the time domain Pilots Data CP/ZP-OFDM: Data TDS-OFDM: (b) Comparison in the frequency domain 3

Application of TDS-OFDM l TDS-OFDM is the key technology of the first-generation DTV broadcasting standard DTMB – IPR-owned: proposed by China (Tsinghua University) in 2006 – Better performance than other standards: DVB-T (EU) , ATSC (USA), ISDB-T (Japan) – Widely deployed: China (Hongkong, Macau), Cuba, Cambodia – ITU approval: approved by ITU as the fourth international DTV broadcasting standard in 2011 4

Challenges of TDS-OFDM l Requirements of next-generation DTV standard – 64 QAM vs. 256 QAM 30% higher spectrum efficiency l Challenges of TDS-OFDM – Mutual interferences Difficult to support 256 QAM in static long-delay channels 5

Theory of Compressive Sensing (CS) l A new sampling theory against Shannon-Nyquist theory – Key point: sparse signal recovery at a rate far lower than traditional Nyquist rate (2 x of the signal bandwidth) – Structured CS 6

Contents 1 Technical Background 2 Proposed Solution 3 Performance Analysis 4 Simulation Results 5 Conclusions 7

Two Properties of Wireless Channels l Sparsity and Inter-Channel Correlation – Wireless channel is sparse in nature – Path delays vary much slower than the path gains – Not considered in conventional TDS-OFDM systems L: Channel length S: Sparsity level S << L Path delay set: 8

TDS-OFDM Based on Simultaneous Multi-Channel Reconstruction Received TS: M interference N (a) Conventional TDS-OFDM IBI-free region: (b) Dual PN padding TDS-OFDM (DPN-OFDM) G (c) Proposed scheme 9

Mathematical problem of structured CS l Mathematical problem of structured CS – Problem: – Solution: – Proposed simultaneous multi-channel reconstruction scheme l Based on classical algorithm called simultaneous orthogonal matching pursuit (SOMP) l Key idea: the specific technical feature of TDS-OFDM is exploited to obtain partial priori of the channel to reduce the complexity 10

Simultaneous Multi-Channel Reconstruction Based on Adaptive SOMP – Step 1: Correlation-Based Channel Priori Acquisition Noise and interference path delays vs. path gains 11

Simultaneous Multi-Channel Reconstruction Based on Adaptive SOMP – Step 2: A-SOMP Based Joint Sparsity Pattern Recovery l Priori is exploited to reduce the complexity l Adaptive to variable channel conditions (channel length, sparsity level, etc. ) l Key difference with SOMP: (S-S 0) instead of S iterations Adaptive SOMP – Step 3: ML-based path gain estimation 12

Contents 1 Technical Background 2 Proposed Solution 3 Performance Analysis 4 Simulation Results 5 Conclusions 13

Performance Analysis (1) l Cramer-Rao lower bound (CRLB) – Conditional PDF – Fisher information matrix – CRLB – Final result : noise level S < G means improved accuracy 14

Performance Analysis (2) l Spectral Efficiency – 10% higher than standard CP-OFDM – 30% higher than conventional TDS-OFDM (64 QAM vs. 256 WAM) 15

Contents 1 Technical Background 2 Proposed Solution 3 Performance Analysis 4 Simulation Results 5 Conclusions 16

Simulation results l Simulation setup – Setup is configured according to the typical wireless broadcasting systems – Signal bandwidth: 7. 56 MHz – Central radio frequency: 770 MHz – FFT size: N=4096 – Guard interval length: M=256 – Modulation schemes: 256 QAM – Channel coding: LDPC code with length of 64800 bits and rate 0. 6 – Channel model: 3 GPP six-tap Vehicular B multipath channel (max. delay of 20 us) 17

Simulation Results (1) l MSE performance comparison in static multipath channel – The proposal outperforms the conventional TDS-OFDM by >5 d. B – The actual MSE performance approaches theoretical CRLB when SNR becomes high 18

Simulation Results (2) l Comparison between A-SOMP and SOMP – A-SOMP requires fewer measurements than SOMP – The MSE approaches theoretical CRLB when G becomes large Size of IBI-free region G 19

Simulation Results (3) l 256 QAM supporting in static long-delay channel – Unlike conventional TDS-OFDM, the proposal can support 256 QAM – The proposed scheme has superior BER performance than DPN-OFDM and CP-OFDM 20

Contents 1 Technical Background 2 Proposed Solution 3 Performance Analysis 4 Simulation Results 5 Conclusions 21

Conclusions l We propose a TDS-OFDM scheme with an improved spectrum efficiency of about 30% for next-generation DTV broadcasting standard l The sparse nature and inter-channel correlation of wireless channels are jointly exploited l The simultaneous multi-channel reconstruction method utilizes multiple IBIfree regions of very small size to reconstruct the wireless channel of high dimension under the newly emerging theory of structured compressive sensing l Not only the obviously improved channel reconstruction accuracy could be achieved, but also the mutually conditional time-domain channel estimation and frequency-domain data detection in conventional TDS-OFDM could be decoupled l The proposed scheme could support 256 QAM in static channel with long delays with a LDPC coded BER performance close to the ideal CSI case l The proposed scheme is directly applicable for unique word single carrier (UW-SC) systems 22

Thank you ! 23