《中興通訊&清華大學：2024年6G近場技術白皮書（英文版）（158頁）.pdf》由會員分享，可在線閱讀，更多相關《中興通訊&清華大學：2024年6G近場技術白皮書（英文版）（158頁）.pdf（158頁珍藏版）》請在三個皮匠報告上搜索。

1、 Weidong Li(),Huazhong University of Science and TechnologyTable of Contents1 Introduction.132 Near-field Application Scenarios.162.1 Near-field in Different Frequency Bands.162.2 Ultra Large Aperture Enabled Near-field.212.3 Integrated Sensing and Communication.252.4 Wireless Positioning.262.5 Simu

2、ltaneous Wireless Information and Power Transfer.272.6 Physical Layer Security.282.7 Multiple Access.292.8 Massive IoT Communications.292.9 On-chip Wireless Communications.303 Fundamental Theories of Near-field.323.1 Near-field Range Partitioning.323.2 Near-field Electromagnetic Physical Effects.353

3、.3 Near-field Degree-of-freedom Theoretical Analysis.413.4 Near-field Performance Analysis and Measurement.434 Channel Measurement and Modeling of Near-field.534.1 Near-field Channel Measurement.534.2 Near-field Channel Simulation.554.3 Near-field Channel Modeling.604.4 Bridging the Gap between Near

4、 and Far-Field Models.705 Transmission Technologies of Near-field.725.1 Near-Field Channel Estimation.725.2 Near-Field Beamforming.775.3 Near-Field Codebook Design.825.4 Near-Field Beam Training.865.5 Near-Field Multiple Access.885.6 Non-Coherent Communication Schemes.915.7 Deployment of Near-Field

5、Communication System.925.8 Standardization.956 Integration of Near-field Technology with Other Technologies.976.1 Near-field Based Positioning.976.2 Integrated Sensing and Communication in Near Field.1066.3 Wireless Power Transfer in Near Field.1136.4 Physical Layer Security in Near-Field.1206.5 Nea

6、r-Field Based OAM.1226.6 Near-Field Based Intelligent Communication.1256.7 Near-field On-chip Wireless Communications.1266.8 Near-field and Material Sensing.1317 Summarization and Prospects.133Reference.13476List of Terms and Abbreviations.15386List of FigureFig.1.1Near-field Application Scenarios.1

7、4Fig.1.2Framework of Near-field Technology.14Fig.2.1IMT-2030 application scenarios and key capability indicators1.16Fig.2.2Full spectrum of high,medium and low bands for the future 6G.17Fig.2.3Multi user near-field communication with beams pointing to each user.18Fig.2.4Non-diffracting beams to circ

8、umvent obstacles.19Fig.2.5An illustration for near field communication feature.For users located in the near fieldregion,spherical wavefronts are utilized for communication compared to plane wavefrontsin the far field region.32.20Fig.2.6RIS assisted near-field application scenarios.22Fig.2.7Near fie

9、ld positioning model38.22Fig.2.8Different architectures of ELAA 39.23Fig.2.9MA for near-field communications and sensing.24Fig.2.10Near field ISAC system37.26Fig.2.11High precision positioning based on near-field effects.27Fig.2.12Schematic diagram of near-field wireless energy transmission.28Fig.2.

10、13Left:Far field secure communication using beam steering.Right:Near field securecommunication using beam focusing.28Fig.2.14Schematic diagram of near-field multiple access.29Fig.2.15Utilizing on-chip and inter-chip communication with antennas.30Fig.2.16Wireless interconnection between chips with di

11、fferent semiconductor materials.31Fig.2.17Single-chip multi-core processors utilizing on-chip wireless communicationtechnology.31Fig.3.1Far-field plane wavefront and near-field spherical wavefront and correspondingphysical space normalized received energy.33Fig.3.2Near-field range for typical commun

12、ication scenarios.34Fig.3.3Near-field electromagnetic radiation system diagram.36Fig.3.4Near-field multi-polarized spherical waves model.38Fig.3.5Near-field tri-polarized channel capacity.38Fig.3.6Near-field beam splitting effect schematic.39Fig.3.7Trend of beam gain with distance.40Fig.3.8Channel c

13、orrelation versus antenna curve.40Fig.3.9Comparison of beamforming gain between UCA and ULA.41Fig.3.10Extra degrees of freedom in the reactive near-field.41Fig.3.11Nyquist sampling under isotropic scattering conditions.42Fig.3.12Singular Values of Near-Field SPD-MIMO.D denotes the transmission dista

14、nce andN is the number of transmit and receive antennas.43Fig.3.13CDF of data rate for co-located and sparse array103.4496Fig.3.14SNRs versus the number of array elements for different models104.45Fig.3.15Near-field beam focusing patterns under different array architectures106107.46Fig.3.16Achievabl

15、e sum rates for modular and co-located arrays versus.46Fig.3.17XL-IRS aided communication system.47Fig.3.18Channel power gain versus the IRS-user distance.47Fig.3.19RIS-assisted Localization Performance.48Fig.3.20Fourier plane wave expansion channel model.49Fig.3.21Fourier plane wave expansion chann

16、el capacity simulation.49Fig.3.22Near-field electromagnetic channel capacity limit.50Fig.3.23HMIMO Communication Application Scenarios.50Fig.4.1Channel measurement platforms in time and frequency domains131142.54Fig.4.2(a)Near-field channel measurements with a virtual array based on VNA.(b)Channelim

17、pulse response on the array elements143.54Fig.4.3Near field simulation in element level.55Fig.4.4The deployment of large antenna array(6GHz,1024elements).56Fig.4.5Absolute delay of a ray per element in BS antenna.56Fig.4.6AoA/AoD/ZoA/ZoD of a ray per element in BS antenna.57Fig.4.7Power gain of a ra

18、y per element in BS antenna.57Fig.4.8Phase of a ray per element in BS antenna.58Fig.4.9The position of antenna array and PEC sphere as well as the incident wave vector.58Fig.4.10The distribution of signal strength gain in antenna array due to H-pol and V-polincident.58Fig.4.11The distribution of pha

19、se in antenna array due to H-pol and V-pol incident.59Fig.4.12The position of antenna array and random scatter as well as the incident wave vector59Fig.4.13The distribution of signal strength gain in antenna array due to V-pol incident.59Fig.4.14The distribution of phase in antenna array due V-pol i

20、ncident.60Fig.4.15Spherical propagation with the SnS characteristic.61Fig.4.16(a)measurement result,(b)generation of channel model.62Fig.4.17Visibility region of the array and user.62Fig.4.18NUSW channel model for spatially discrete antennas.64Fig.4.19Greens function-based channel model for CAP ante

21、nnas.65Fig.4.20Modeling of near-field multi-polarized spherical waves.65Fig.4.21Multi-polarized channel capacity.66Fig.4.22XL-MIMO hybrid near and far field propagation environment.67Fig.4.23Illustration of the planar and spherical wavefronts in the RIS-enabled channel model68Fig.4.24Illustration of

22、 the sub-array partition model.68Fig.4.25Comparisons of the channel modeling accuracy based on the sub-array partition-basedalgorithm and that based on the far-field planar wavefront model under different motiontime instants and different RIS units.70Fig.4.26Comparison of near and far-field models f

23、or different Tx-Rx distances 173.71Fig.5.1The energy spread effect in the angle-domain.73Fig.5.2Dictionary coherence comparison.74106Fig.5.3Space Partition Based on Joint Angular-Polar Domain Transform.74Fig.5.4MRDN-based channel estimation scheme.76Fig.5.5P-MRDN-based channel estimation scheme.76Fi

24、g.5.6RDN、CMAM and ASPP-RDN system models.76Fig.5.7Far-field beamforming and Near-field beamforming.78Fig.5.8Fully-Connected Delay-Phase Hybrid Beamforming.79Fig.5.9Partially-Connected Delay-Phase Hybrid Beamforming.79Fig.5.10Serially-Connected Delay-Phase Hybrid Beamforming.79Fig.5.11Extremely large

25、-scale MIMO system with CPU and LPU collaborative processing.81Fig.5.12PAA-RIS dual beamforming scheme.81Fig.5.13Rate heatmap comparison(from left to right:near field beam focusing,far fieldbeamforming,the proposed variable beamwidth near field beamforming).82Fig.5.14Illustration of FRFT codeword qu

26、antization performance.83Fig.5.15Angle-displaced near-field codebook design method.83Fig.5.16Principle and phase distribution of far-and near-field codebooks.84Fig.5.17Codeword coverage discretization in the near-far field codebook.85Fig.5.18Neural network structure for near-field beam training.88Fi

27、g.5.19Comparison between far-field SDMA and near-field LDMA.89Fig.5.20Illustration of near-field NOMA designs.90Fig.5.21Deployment scenarios of XL-arrays(taking RIS as an example).94Fig.5.22A hybrid communication architecture based on near-field relays.95Fig.6.1Near-field signal model and far-field

28、signal model.97Fig.6.2Near-field positioning and attitude sensing.98Fig.6.3User positioning with sub-arrays.99Fig.6.4The received phase of a spherical wave on a large array and its approximation withplanar waves on 5 subarrays.100Fig.6.5Schematic diagram of two-dimensional DoA estimation based on RI

29、S and non-uniformtime modulation 241.100Fig.6.6RIS-assisted THz multi-user positioning systems.102Fig.6.7RMSE versus the number of RIS elements.102Fig.6.8Localization system coordinate.103Fig.6.9PEB as a function of distance to the RIS using random,directional and positional RISphase profiles.103Fig

30、.6.10Data Rate versus RIS distance.104Fig.6.11Schematic diagram of near-field beam squint.105Fig.6.12Schematic diagram of near-field controllable beam squint.105Fig.6.13Near-field radar sensing with XL-MIMO.107Fig.6.14CRB of angle for monostatic sensing.108Fig.6.15CRB of range for bistatic sensing.1

31、09Fig.6.16Far-field Velocity Sensing.109Fig.6.17Near-field Velocity Sensing.110116Fig.6.18Communication-assisted near-field sensing and sensing-assisted near-fieldcommunication.111Fig.6.19Experimental environment for near-field sensing,and measurement results for sensingaccuracy in terms of signal b

32、andwidth 259.113Fig.6.20Adaptive intelligent near-field charging system based on programmable metasurface267.115Fig.6.21Schematic diagram of multi-target WPT system based on quasi-Bessel beams 249.115Fig.6.22Block diagram of wireless energy harvesting system.116Fig.6.23Schematic diagram of the recti

33、fying metasurface.116Fig.6.24SWIPT systems based on(a)frequency diversity and(b)polarization diversity.117Fig.6.25Near-Field SWIPT.118Fig.6.26Holographic WPT 282.119Fig.6.27Holographic SWIPT 282.119Fig.6.28Near-Field PLS.121Fig.6.29Beam diffraction in near-field.122Fig.6.30Comparison of the electric

34、 field between(a)conventional OAM beam and(b)non-diffraction Bessel vortex beam 291.124Fig.6.31Schematic diagram of(a)full aperture sampling and receiving method and(b)partialaperture sampling and receiving method 294.124Fig.6.32An illustration of the near-field based semantic communication system.1

35、25Fig.6.33An illustration of the near-field based federated learning framework.125Fig.6.34Near-field intelligent beamforming(left)and performance comparison(right).126Fig.6.35Commonly used on-chip communication system block diagram295.127Fig.6.36Layout representation for the intra-chip communication

36、 arrangement299.127Fig.6.37Cross-section of the on-chip top-hat antenna 303.128Fig.6.38On-chip antenna based on GaN technology 303.128Fig.6.39(a)Typical coil array structure,(b)Coil array structure with a shielding pattern,(c)Proposed zigzag-shaped coil array for wireless chip-to-chip communication.

37、130Fig.6.40Illustration of wireless in-plane/out-of-plane intra-/interchip communications utilizingTGV-integrated antennas in 3D System-in-Packaging(SiP).131Fig.6.41Sensing scenario.13112613/1561 1 IntroductionAs the commercialization of 5G wireless networks gains momentum,there is a growing emphasi

38、son exploratory research into the upcoming 6G wireless networks.This era of technologicaladvancement sees 6G networks characterized by a more visionary and performance-driven ethoscompared to their predecessors.Traditional wireless networks,spanning from 1G to 5G,predominantlyoperate within a spectr

39、um below 6 GHz,often below 3 GHz.Due to the physical constraints on antennaarrays and the proportional relationship between element spacing and wavelength,these spectrumbands necessitate the use of relatively small numbers of antennas.Consequently,the combined effectof these lower-dimensional antenn

40、a arrays and lower frequency bands confines the range of wirelessnear-field communication to mere meters or even centimeters,thereby shaping the design of thesesystems around far-field assumptions.However,the transition to 6G networks is marked by the adoption of larger antenna apertures andhigher f

41、requency bands,such as new mid frequency,millimeter wave,and terahertz bands,whichaccentuate near-field characteristics,as shown in Table 1.1.The integration of emerging technologiessuch as Reconfigurable Intelligent Surface(RIS)123,extremely large aperture arrays(ELAA)4,movable antenna(MA)5and cell

42、-free networks6is expected to amplify the prevalence of thenear-field scenario in future wireless networks.This shift challenges traditional far-field plane waveassumptions and underscores the need to rethink strategies for spatial resource utilization7.Whileconventional systems have effectively exp

43、loited far-field spatial resources,the exploration andutilization of near-field spatial resources in 6G networks promise to introduce novel physicaldimensions to wireless communication systems.This shift towards near-field regions in 6G networkscatalyzes a new wave of research in near-field technolo

44、gy paradigms.Table 1.1 Near field range of typical scenarios(Rayleigh distance)Df2.6 GHz(low band)7 GHz(Mid band)28 GHz(mmWave band)220 GHz(THz band)0.5m4 m12 m483721.6m60 m119 m476/3.0m210 m420 m/In the realm of near-field technology,the nuanced propagation behaviors of electromagneticwaves necessi

45、tate a departure from plane wave approximations to embrace spherical wave treatments.This paradigm shift brings forth a plethora of hitherto overlooked electromagnetic phenomena,including spatial non-stationarity,tri-polarization,evanescent waves,and the capacity for near-fieldfocusing,all of which

46、pose challenges to the efficacy of traditional communication algorithms withinthe context of 6Gs near-field environment.Harnessing near-field effects holds the promise of better realizing a broader range of applicationscenarios and key performance indicators outlined in IMT-2030.This article compreh

47、ensively explorespotential application scenarios based on near-field technology,as shown in Fig.1.1.14/156Fig.1.1 Near-field Application ScenariosThrough meticulous examination of both theoretical foundations and technological advancements,we have meticulously crafted a preliminary framework for nea

48、r-field technology,as illustrated in Fig.1.2.Our discourse commences with an exploration of the definition of near-field in electromagnetictheory,tracing the origins of near-field electromagnetic effects and their ramifications on existingcommunication systems.Drawing upon an extensive body of liter

49、ature,we provide an encompassingsynthesis of near-field effects on communication system design and performance,with particularemphasis on degrees of freedom and communication capacity.Fig.1.2 Framework of Near-field Technology15/156A nuanced comprehension of near-field channel characteristics and mo

50、dels is pivotal forcommunication system design and evaluation.Thus,our article underscores the imperative forexhaustive channel measurements and precise channel characterization.Furthermore,we delve intonear-field transmission technologies,encompassing facets such as channel estimation,beamforming,c

51、odebook design,beam training,multiple access technology,system architecture,deploymentconsiderations,and implications for standardization.Additionally,we explore the convergence ofnear-field technology with other domains,including positioning,wireless power transfer,physical layersecurity,orbital an

52、gular momentum-based near-field,AI-driven communication,and Near-fieldOn-chip Wireless Communications.Despite notable advancements in the research on near-field propagation characteristics,a dearth ofliterature offers a systematic synthesis of near-field technology.Therefore,this article endeavors t

53、obridge this gap by furnishing an all-encompassing summary of near-field technologys applicationscenarios,fundamental theories,channel measurement and modeling methodologies,transmissiontechnologies,and integration with allied fields.Our overarching aim is to cultivate and propel theadvancement of n

54、ear-field technology research.16/1562 2 Near-fieldApplication ScenariosIn November 2023,the International Telecommunication Unions 5D Working Group on WirelessCommunications(ITU-R WP5D)released a framework and overall goal proposal for the developmentof IMT towards 2030 and the future,proposing typi

55、cal 6G scenarios and capability indicator systems,as shown in Fig.2.1 18.6G scenarios include immersive communication,ultra large-scaleconnections,extremely high reliability and low latency,artificial intelligence and communication,integration of perception and communication,ubiquitous connections,e

56、tc.The key 6G capabilityindicators include 9 enhanced 5G capabilities and 6 new capability dimensions,including peak datarate,user experienced data rate,spectrum efficiency,regional traffic,connection density,mobility,latency,reliability,security privacy elasticity,coverage,sensing related indicator

57、s,applicable AIrelated indicators,sustainability and positioning 9.6G is expected to continue the efforts of 5Gadvanced in terms of keeping enhancing connection experiences for mobile users and enabling morevertical industries 10.Fig.2.1 IMT-2030 application scenarios and key capability indicators1T

58、o meet the spectral efficiency requirements of IMT-2030,it is necessary to further explore theapplication potential of higher frequency bands and larger scale arrays.Simultaneously,the ultra large-scale arrays in higher frequency bands will bring near-field effects.The so-called near-field effectref

59、ers to the situation where,under certain distance conditions,the assumption of electromagneticwaves as plane waves in the far-field no longer holds and needs to be modeled as spherical wave.Thesphericalwavefrontcarriesnotonlyangleinformation,butalsodepthinformation.Theelectromagnetic beam focusing h

60、appens simultaneously in both the angle and depth domains,formingnear-field beam focusing 11.By utilizing the near-field effect,more application scenarios and keyperformance indicators of IMT-2030 can be better achieved,such as the integration of perception andcommunication,positioning,security,mobi

61、lity,etc.This section will explain the application scenariosof near field based on the above analysis.2.1 Near-field in Different Frequency BandsThe expansion of bandwidth and the increase of antennas will bring greater capacity and higherspectral efficiency to wireless communication systems.Typical

62、 2G,3G,4G,and 5G communicationsystems use bandwidths of 0.2 MHz,5 MHz,20 MHz,and 100 MHz,respectively,and largerbandwidths will be required for 6G in the future.17/156In May 2023,Chinas Ministry of Industry and Information Technology(MIIT)issued a newversion of the Regulations of the Peoples Republi

63、c of China on the Division of Radio Frequencies(MIIT Decree No.62),which is the first in the world to use all or part of the bandwidth of the6425-7125 MHz frequency band,totaling 700 MHz,for the 5G-A/6G system 12.In December of thesame year,the International Telecommunication Union(ITU)held the Worl

64、d RadiocommunicationConference 2023(WRC-23)in Dubai,UAE,and completed a new round of revisions to the RadioRules,newly dividing 6G spectrum resources in the mid-band of 6425-7125 MHz with a total of 700MHz of bandwidth for most of the countries in the world 13.In December 2023,the internationalstand

65、ardization organization,3GPP,held its Edinburgh,UK meeting,in which the first projects forRel-19,the second version of the 5G-Advanced standard,including eight areas such as channelmodeling studies for the new 7-24 GHz spectrum are established 14.Compared to the sub-6 GHz low-frequency band,which is

66、 widely used in 5G,and thehigh-frequency bands,such as millimeter wave and terahertz,which may be used in 6G in the future,the mid-frequency band,which combines the advantages of both coverage and capacity,is of greatvalue for the wide-area high-capacity coverage of 6G,and it is expected to be one o

67、f the fundamentalfrequency bands for 6G,as shown in Fig.2.2.Fig.2.2 Full spectrum of high,medium and low bands for the future 6GIn the future,6G high,medium,and low frequency bands are likely to face near-fieldcommunication.6G Technologies,a research report published by the 6G Alliance in June 2022,

68、alsoclearly points out the necessity of researching near-field in 6G high,medium,and low frequency bands15.2.1.1High Frequency Band TransmissionMillimeter wave(mmWave)and terahertz(THz)wireless communication can utilize largeavailable bandwidth to improve data transmission rates,making it one of the

69、 key technologies for thenext generation of communication systems 161718.In order to compensate for the path loss ofhigh-frequency transmission,base stations(BS)operating on these frequency bands will be equippedwith large-scale antenna arrays.The application of large-scale antenna arrays will incre

70、ase thepossibility that the users in high-frequency communication falling into the near-field region,whiletraditional wireless systems typically operating in the far-field range.Under millimeter wave andterahertz conditions,the near-field distance of relatively small antennas/surfaces can also reach

71、 severaltens of meters.For example,the near-field distance of a uniform linear array with 128 antennasworking at 300 GHz would be 65 meters,which covers a relatively large area.Namely,the far-fieldplane wave assumption on electromagnetic fields is no longer applicable at actual communications18/156d

72、istances.Thus,a near-field model of spherical waves should be used.The management of sphericalwavefront can be transformed into flexible beamforming ability.For example,utilizing the sphericalwavefront can focus the electromagnetic waves into a spot rather than traditional beam steering underfar-fie

73、ld condition,which is referred as the concept of beam focusing in recent literature19.Beamfocusing can support multiple orthogonal links even at similar angles.The ability to focus beams in large-scale multiple input multiple output(MIMO)systems largelydepends on the signal processing capability of

74、the antenna array,with different processing capabilitiesin different architectures.The most flexible solution for a given radiation element array is full-digitalarchitecture,where each antenna element is connected to a dedicated radio frequency(RF)chain.Under this architecture,the transceiver can si

75、multaneously control infinite beams in multiple directions,greatly improving spatial flexibility.However,when deploying large-scale arrays in 5G and moreadvancedcommunicationsystems,implementingfull-digitalarchitecturesbecomesextremelychallenging due to increased costs and power consumption.To allev

76、iate this,large-scale MIMOcommunications adopt a hybrid analog-digital architecture.This hybrid architecture combines lowdimensional digital processing and high-dimensional analog precoding,typically achieved throughphase shifter interconnections,resulting in fewer RF links than antenna components.A

77、nother emergingtechnology for effectively implementing large-scale arrays is dynamic meta-surface antennas,whichcan programmatically control the transmit/receive beam patterns,provide advanced analog signalprocessing capabilities,and naturally achieve frequency chain reduction without the use of ded

78、icatedanalog circuits.This also densifies antenna components,thereby improving focusing performance.Reference20 explores multi-user communications in near-field utilizing various antenna architectures,including all digital arrays,hybrid architectures based on phase shifters,and dynamic meta-surfacea

79、ntennas,as well as the impact on downlink multi-user systems when forming focused beams.Fig.2.3 Multi user near-field communication with beams pointing to each user(a)towards each user in three-dimensional space;(b)beam steering in far-field,leads to interferencebetween users at the same angle;(c)be

80、am focusing in near-field,with minimal interference20In addition,new types of wavefronts become available beyond spherical waves when operating inthe near field21.By wavefront,we refer to the imaginary surface representing all points in a wavethat are in the same phase at a given time.Among others,t

81、he use of Bessel beams has been recentlyproposed2223.Bessel beams are no-diverging beams that focus the signal along a line.To generatean ideal Bessel beam,one would need an infinite aperture(i.e.,an infinite lens or antenna array orreflect-array).When using a finite aperture,the Bessel beam only ex

82、ist until a maximum distance19/156delimited in the near-field.Focusing the power along a line,instead than on a point,can drasticallyreduce the amount of channel state information to ensure the reliable transmission of information.Moreover,Bessel beams are self-healing,i.e.,even when they are partia

83、lly blocked,the signal isregenerated to the original level after the obstacle 2425.This can be leveraged to overcomeblockage,which is one of the main problems for high-frequency systems.Another type ofnon-conventional wavefronts that can be generated in the near field are Airy beams 26.These are als

84、onon-diffracting beams that,in this case,focus along a curving line.This allows for example tocircumvent obstacles(See Fig.2.4).Fig.2.4Non-diffracting beams to circumvent obstaclesa)Illustration of a Bessel beam,focusing a long a line in the near field;b)Phase distribution of a Besselbeam;c)Bessel b

85、eam regeneration after an obstacle;d)Phase distribution of anAiry beam;e)AcurvingAiry beam.Traditionally,Bessel beams and Airy beams have been generated at optical frequencies andutilizing different types of lenses.For example,axicons are utilized to generate Bessel beams.Nevertheless,these wavefron

86、ts can be generated utilizing arrays,reflect-arrays and meta-surfaces27with a large number of elements2829 at least phase control per element.Operating with new typesof wavefronts drastically changes many well-known concepts 30,including the study of interferenceacross different type of beams,channe

87、l state information estimation,joint ultrabroadband waveformand wavefront design 28,and even physical layer security31.2.1.2Mid Frequency Band TransmissionIn September 2020,the Third Generation Partner Program(3GPP)completed the approval of the 6GHz licensed frequency band,initiated the standardizat

88、ion of mid frequency band RF.Then,3GPPsuccessfully completed the standardization of the 6425-7125 MHz licensed frequency band in20/156September 2022.Subsequently in December 2023,3GPP further listed channel modeling in the 7-24GHz mid frequency band as one of the earliest standard topics in the seco

89、nd standard version of5G-Advance,continuing promoting the standardization process of the mid frequency band.At the sametime,in May 2023,the Ministry of Industry and Information Technology of China issued a new versionof the Regulations on the Classification of Radio Frequency in the Peoples Republic

90、 of China,whichclearly divides the 6 GHz mid frequency band for 5G/6G mobile communication systems.The 6 GHzfrequency band has a continuous large bandwidth of 1200 MHz,which has lower propagation path lossand stronger coverage ability compared to the high-frequency band.It has both coverage and capa

91、cityadvantages and can be used for wide area high-capacity coverage in 6G communication.Therefore,themid frequency spectrum resources will become one of the important alternative frequency bands for 6Gcommunication.Compared to the sub-6 GHz frequency band of 5G communication,the increase in communic

92、ationfrequency leads to weaker coverage due to large propagation losses in materials and smaller antennasin mid frequency centimeter wave communication.In order to compensate for the high path loss duringthe propagation of intermediate frequency signals,the antenna size of intermediate frequency bas

93、estations needs to be further increased.With the increase of communication frequency band and basestation antenna array aperture,the near-field range of intermediate frequency communication will alsobe significantly expanded.Taking the 7 GHz communication system and base station with 1.6-meterantenn

94、a array as an example,its near-field range exceeds 100 meters.In the formal project of the 3GPPintermediate frequency channel modeling standard,near-field characteristics and spatial non-stationarycharacteristics are considered as new characteristics of mid frequency channels,becoming importantconsi

95、derations for improving the 3GPP channel model 14.Therefore,near-field spherical wavecommunication will become an important scenario for mid frequency transmission.Due to the shorter wavelength of mid frequency centimeter waves and the further increase inantenna size,mid frequency system is more lik

96、ely to form high-resolution narrow spatial beams,therefore achieving higher spatial degrees of freedom.Typical application scenarios for this bandinclude single user multi-stream or higher-order multi-user multiplexing scenarios.The modeling andmeasurement of near-field spherical wave channels in th

97、e mid frequency range will provide a channelmodel foundation for mid frequency communication.The near-field spherical wave propagation modelis expected to provide additional single user transmission spatial freedom.The focusing characteristicsof near-field spherical waves can also be used for high-o

98、rder multi-user multiplexing methods,furtherimproving the throughput performance of 6G new mid frequency communication systems.Fig.2.5An illustration for near field communication feature.For users located in the near field region,spherical wavefronts are utilized for communication compared to plane

99、wavefronts in the far fieldregion.322.1.3Low Frequency Band TransmissionThe low frequency band(FR1,Sub-6GHz)defines the baseline coverage range of cellularnetworks.While expanding to higher frequency bands,6G will also fully utilize the advantages of FR1frequency band for wide coverage and deep pene

100、tration to improve spectrum efficiency and breakthrough bandwidth bottlenecks.Large scale MIMO can be used in the low frequency range to improvethe spectral and energy efficiency of 6G systems while ensuring wide coverage.If traditional large-scale MIMO is deployed in the low frequency range,it will

101、 face limitations onantenna size due to tower or base station deployment.Modular or distributed large-scale MIMO,aswell as meta-surface antennas,are expected to overcome size limitations and reduce the requirement ofhalf wavelength distance between antenna units through compact antenna arrays.On the

102、 other hand,traditional cell-based deployment strategies pose challenges such as feasibility,processing,andarchitectural complexity.Therefore,large-scale MIMO in the low frequency range may adopt multipanel,multi transceiver nodes,non-cellular,and irregular large-scale distributed network deployment

103、.In this scenario,further research is needed on distributed deployment strategies,the potential demandfor new channel models from non-uniform antenna panels,large antenna arrays,and near-field effectswhen users may approach access points.Also,research that focuses on exploring efficient referencesig

104、nal designs,channel acquisition frameworks for far-field and near-field channels,further evaluatingthe potential of artificial intelligence in channel acquisition would be valuable.2.2 Ultra Large Aperture Enabled Near-field2.2.1RIS Enabled Near-fieldReconfigurable Intelligent Surface(RIS)is conside

105、red as one of the key potential technologies in 6G,consisting of a large number of low-cost reconfigurable units 33.Deploying RIS in wireless networks can effectively adjust the wireless channel between transmitters and receivers,thereby improving communication quality and coverage range,as already

106、demonstrated by field trials 34.One of the typical applications of RIS technology is to obtain sufficient beamforming gain through hundreds or even thousands of components for coverage blinding in millimeter wave and terahertz communications.The larger RIS array and higher operating frequency furthe

107、r expand the near-field area of RIS assisted communication links 35.RIS is typically used to establish a direct connection channel between transmitters/receivers.In the far-field region,the rank of the channel is usually small,which restricts the spatial multiplexing gain of the channel.On the other

108、 hand,due to the nonlinear changes in signal amplitude and phase caused by spherical waves,near-field channels have better rank conditions,which can effectively improve the multiplexing gain and spatial freedom of the system 36.When users are located in the radiation near-field region,even if multip

109、le users are at the same radiation angle,different near-field codebooks can be configured on the intelligent meta-surface to reduce co-channel interference through beam focusing,supporting multiple coexisting orthogonal links to achieve space division multiple access 37,as shown in Fig.2.6.Similarly

110、,the degrees of freedom provided by the spherical wavefront and the near-field radiation wave carrying both angle and distance information further enhances the accuracy of wireless positioning services and perception,as shown in Fig.2.7.On the other hand,this also means that the spatial non-stationa

111、rity of the channel is21/15622/156intensified,which will bring challenges to channel estimation,codebook design,beam trainingcomplexity,mobility management,signaling design,and other aspects.Fig.2.6 RIS assisted near-field application scenarios.Fig.2.7 Near field positioning model382.2.2ELAA Enabled

112、 Near-fieldExtremely Large Aperture Array(ELAA)is essential to the candidate technologies for 6G such asExtremely Large-Scale MIMO(XL-MIMO).Compared to 5G massive MIMO,ELAA for 6G not onlymeans a sharp increase in the number of antennas but also results in a fundamental change of theelectromagnetic(

113、EM)characteristics.With the significant increase of the antenna number and carrierfrequency in future 6G systems,the near-field region of ELAA will expand by orders of magnitude.The two commonly used ELAA architectures are co-located and distributed ELAAs,as shown in Fig.2.8(a)and(b).The antenna ele

114、ments of co-located ELAA are typically separated by half wavelength,and its physical dimension is limited by the continuous platform 39.By contrast,distributed ELAA isan architecture that antennas are widely distributed over a vast geographical region with multipleseparated sites,which are interconn

115、ected by the backhaul/fronthaul links,so as to perform joint signalprocessing.However,distributed ELAA,e.g.,cell-free ELAA,usually requires the sophisticated sitecoordination and high backhaul/fronthaul capacity.In order to complement for existing ELAA architectures,the works 4041 propose a novelmod

116、ular ELAA architecture.As illustrated in Fig.2.8(c),the antenna elements of modular ELAA areregularly mounted on a shared platform in a modular manner.Each module is comprised of amoderate/flexible number of array antennas with the inter-element distance typically in the order of the23/156signal wav

117、elength,while different modules are separated by the relatively large inter-module distance,so as to enable conformal capability with the deployment structure in practice.For example,themodular ELAA with interlaced modules can be embedded into the discontinuous wall spaced bywindows,like facade circ

118、umstances of shopping malls,factories or office buildings.Compared toco-located ELAA with the same number of antenna elements,modular ELAA not only has thecharacteristic of flexible deployment,but also a higher spatial resolution due to the larger physicaldimension.However,since the inter-module dis

119、tance is much larger than half wavelength,modularELAA will lead to the undesired grating lobes.On the other hand,different from the distributed ELAAarchitecture,modular ELAA typically performs joint signal processing,without having to exchange orcoordinate sophisticated inter-site information,which

120、may ease the requirement of synchronization andreduce hardware cost associated with the backhaul/fronthaul links for distributed ELAA.Uniform sparse ELAA is array architecture where the inter-element spacing is larger thanhalf-wavelength,as illustrated in Fig.2.8(d),which is a special case of modula

121、r ELAA.In general,uniform sparse ELAA results in a narrower main lobe due to the higher spatial resolution,which canprovide a significant interference suppression gain in scenarios with densely located users 42.Similarto modular ELAA,uniform sparse ELAA gives rise to the undesired grating lobes,due

122、to theinter-element spacing much larger than half wavelength.It is worth mentioning that the above four array architectures are suitable for different applicationscenarios.For example,co-located,modular and uniform sparse ELAAs can all be used to supportcellular hotspot communications,while modular

123、and uniform sparse ELAAs achieve a highertransmission rate in scenarios with densely located users.Besides,distributed ELAA is able to providea better communication service for geographically widely distributed users.Thus,the above fourarchitectures complement each other,and the choice of appropriat

124、e ELAA architecture depends on theactual application scenario.(a)co-located ELAA(b)Distributed ELAA(c)Modular ELAA(d)Uniform sparse ELAAFig.2.8 Different architectures of ELAA392.2.3Cell-Free Enabled Near-fieldUnlike the classic cellular communication architecture,the Cell-Free communication archite

125、ctureachieves a user centered communication paradigm by deploying a large number of access nodes in adistributedmanner,effectivelyovercomingintercellinterference,avoidingcommunicationinterruptions,and further improving the performance of next-generation 6G mobile communication.Based on the Cell-Free

126、 communication architecture,the equivalent array aperture is significantlyexpanded due to the distributed deployment of multiple arrays,and the near-field spherical wave effectis more significant.Meanwhile,due to the denser distribution of access nodes and shortercommunication distances,users will h

127、ave a higher probability of being in the near-field range.Inaddition,due to the collaborative nature of non-cellular communication architectures,users may beserved simultaneously by multiple access nodes with different antenna sizes and distances,which maybe located in the far-field or near-field ra

128、nge of different nodes,facing more complex mixed far-fieldand near-field communication scenarios.Therefore,cellular free near-field communication will be oneof the important application scenarios for future 6G.The modeling of near-field spherical wave channels can provide a model foundation for cell

129、ularfree communication systems.Due to its significant near-field spherical wave effect,considering thenear-field spherical wave property can further improve the optimization accuracy of access nodes incellular free architectures.At the same time,beamforming methods that are compatible with near-fiel

130、dspherical waves and far-field plane waves,efficient far-field cellular free communication channelestimation,and beam training schemes can better adapt to near-field communication scenarios,furtherimproving the performance of cellular free communication systems.2.2.4MA Enabled Near-FieldFig.2.9 MA f

131、or near-field communications and sensingMA technology has recently been introduced in wireless communication systems to control the local movement(in position and/or rotation)of antennas at the Tx/Rx for improving wireless channel conditions and communication performance 43.There are various practic

132、al methods that can be used to enable antenna movement,such as mechanical motors,microelectromechanical system(MEMS),etc.Due to their flexible movement capability,MAs can fully exploit the wireless channel spatial variation.24/156For example,they can significantly enhance the spatial diversity perfo

133、rmance,in terms of receiver signal power improvement and interference mitigation,as compared to conventional fixed-position antennas 4445.Besides,for multi-MA aided MIMO and/or multiuser communication systems,the channel matrices can be reshaped by antenna position optimization to increase the spati

134、al multiplexing gain and thus the wireless channel capacity 464748.Moreover,by integrating multiple MAs into an array,more flexible beamforming can be realized by jointly designing the array geometry and beamforming vector 4950.Since the effective array aperture scales with the size of the antenna m

135、oving region,enlarging the antenna moving region expands the near-field region of the Tx/Rx for communication as well as sensing,as shown in Fig.2.9.Different from the ELAA which requires an extremely large number of antenna elements and radio frequency(RF)frontends,the number of MAs is moderate and

136、 can be kept constant even with the increasing moving region size.Thus,MAs can help reduce the hardware cost and RF power consumption as compared to ELAA.The performance advantages of MA systems,such as higher spatial diversity,enhanced multiplexing gain,and more flexible beamforming,become more app

137、ealing in 6G near-field communications because the spherical wave-based model renders more substantial channel variation in the spatial domain.Furthermore,distributed MAs can be seamlessly integrated into cell-free communication systems,providing additional degrees of freedom in antenna position and

138、/or rotation for improving the performance of 6G networks.In wireless sensing and ISAC applications for 6G,the MA systems can effectively enlarge the antenna aperture such that the angular/ranging accuracy is increased manifoldly.For sufficiently large antenna moving regions,the MA-aided systems can

139、 realize super-resolution for near-field sensing.In summary,MAs opened up a new direction for research in 6G near-field communication and/or sensing.More collective efforts in theoretical research,technical exploration,system design,experimental verification,and standardization activities are requir

140、ed to unleash the full potential of MAs in future 6G networks.2.3 Integrated Sensing and CommunicationIn addition to high-capacity communication,the next generation of wireless networks also has the potential to achieve high-precision perception.Therefore,the integrated sensing and communication(ISA

141、C)technology has also attracted widespread research interest in academia and industry 51.Compared with traditional wireless positioning and channel estimation,wireless perception relies on the echo signal reflected by passive targets,rather than the pilot signal sent by active devices.Currently,many

142、 existing modulation waveforms have been proven to be applicable to wireless sensing,such as orthogonal frequency division multiplexing(OFDM)and orthogonal time frequency space(OTFS),indicating that sensing functions can be seamlessly integrated into existing wireless communication networks 5253.Bes

143、ides,novel dual-functional waveform design strategies have been developed to balance the communication and sensing performances under different application scenarios,e.g.,ISAC at mmWave/THz frequencies or communication/sensing-centric ISAC services.In far-field sensing,increasing the size of the ant

144、enna array often only improves the resolution of angle estimation,while the resolution of distance and velocity mainly depends on signal bandwidth and perception duration.However,in the near-field region,the propagation of spherical waves allows large-scale antenna arrays able to estimate the distan

145、ce and movement speed between objects.On the one hand,even within a limited bandwidth,near-field channels can still effectively contain distance25/15626/156information,improving the resolution of distance estimation in narrowband systems.On the other hand,the estimation of target velocity depends on

146、 the estimation of Doppler frequency.Compared withfar-field sensing,near-field sensing may have significantly different Doppler frequencies whenobserving two antennas in a large-scale antenna array from different directions,which can enhance theestimation of object movement speed 3754,as shown in Fi

147、g.2.10 Based on the above discussion,near-field effects have the potential to promote high-precision perception in situations wheretime-frequency resources are limited.Therefore,near-field synesthesia integration is a highlypromising technology.Fig.2.10 Near field ISAC system372.4 Wireless Positioni

148、ngIn traditional far-field communication systems,the angle and distance information of the targetrelative to the receiving point is mainly obtained by estimating the arrival angle and time of the signalat the target based on the assumption of plane waves 55.The far-field communication system needsto

149、 deploy multiple receiving points as positioning anchors to estimate the three-dimensionalcoordinates of the target based on the angle and distance information of multiple anchors.In order toobtain more accurate angle and distance information,far-field communication systems usually need toconfigure

150、measurement signals with larger bandwidth.In addition to the use of distance and angle,theuse of the characteristics of the received signal as a fingerprint for localization is also a commonmethod of localization and has been studied in far-field communications 56.In the near field,basedon the spher

151、ical wave model,the arrival angles of signals from antenna units in different regions of theantenna array at the target are different.By utilizing the signal transmission characteristics of beamconvergence,near-field communication systems locate targets through the differences in channelangles in di

152、fferent areas of the antenna array,thereby reducing the demand for measurement signalbandwidth 56.Meanwhile,the deployment of large-scale antenna arrays is beneficial for furtherenhancing angular resolution and providing additional distance resolution in the near-field region,which is conducive to a

153、chieving high-precision positioning in 6G mobile communication 57.Fig.2.11 High precision positioning based on near-field effectsHigh precision positioning services can be provided in the near-field through various forms such asExtremely Large Aperture Array(ELAA),Reconfigurable Intelligent Surface(

154、RIS),and distributedMIMO(D-MIMO).The positioning process of near-field communication systems is different from traditional far-field communication systems in terms of signal system,channel model,and positioning principle.The two belong to heterogeneous positioning networks.Therefore,heterogeneous po

155、sitioning network fusion algorithms are needed between far-field and near-field communication systems to ensure seamless positioning services 5960.The fusion of heterogeneous positioning networks relies on the implementation of positioning accuracy estimation algorithms 6162.For regional positioning

156、 systems,including near-field communication systems,positioning accuracy algorithms can evolve into availability estimation 63 to support two different modes of interoperability between heterogeneous positioning systems:soft fusion and hard switching 64.2.5 Simultaneous Wireless Information and Powe

157、r TransferIn near-field communication,a highly directional point beam can be achieved,which concentrates the target area of the beam near the target device,thereby concentrating the energy of the RF signal to the energy collection node of the Internet of Things device.By utilizing the large number o

158、f antennas and high-precision position information,the efficiency of wireless energy transmission can be significantly improved,reducing energy waste during the transmission process.The near-field beam focusing characteristics limit the spread of energy to undesired location,but does not affect the

159、efficiency on its own.In indoor scenarios or scenarios where the size of base station antennas is limited,wireless communication systems can use intelligent meta-surfaces to construct near-field channels and gather signal energy from home base stations to energy harvesting nodes.In addition,in the n

160、ear-field communication system,the super large antenna array can obtain higher spatial resolution in the near-field range based on the wireless channel of spherical wave model,so that the base station can support higher density simultaneous wireless information and power transfer(SWIPT)terminals.SWI

161、PT allows devices to harvest energy from RF waves and convert it to electrical energy,storing that27/15628/156energy into the devices battery,maximizing the devices lifespan and representing a new solution tolimited energy6566.Fig.2.12 Schematic diagram of near-field wireless energy transmission2.6

162、Physical Layer SecurityDue to the natural broadcasting and mobile characteristics of wireless communication,thecommunication of legitimate users in the network is easily eavesdropped and attacked by illegal users,and secure transmission has always been an important issue in wireless communication.In

163、 far-fieldcommunication,if the eavesdropper is in the same direction as the legitimate user,especially when theeavesdropper is closer to the base station,secure transmission will be difficult to achieve.Unlike thedirectional focusing of beamforming in far-field communication,in near-field communicat

164、ion assistedby ultra large arrays,the beams formed by base stations have strong positional focusing 67.Thisproperty allows the energy of the transmitted signal to concentrate at the location of legitimate usersrather than just in their direction,effectively reducing information leakage at the locati

165、on ofeavesdropping users and improving the systems secure channel capacity.By optimizing the beamfocusing design of the base station,the potential of near-field communication in enhancing physicallayer security can be fully explored.Fig.2.13 Left:Far field secure communication using beam steering.Ri

166、ght:Near field secure communicationusing beam focusing.29/1562.7 Multiple AccessMultiple access technologies leverage time,frequency,code,and space domains in order toachieve the efficient utilization of transmission resources,which is the key to improving thetransmission rate of the next-generation

167、 wireless network.In the current 5G massive MIMO system,spatial division multiple access(SDMA)utilizes orthogonal resources in the angular domain todistinguish different users;non-orthogonal multiple access(NOMA)further allows multiple users toreuse the same resource block and mitigate the inter-use

168、r interference through the power or codedomain;unsourced random access introduces a competition mechanism of access resources(such ascodewords),which saves the resource overhead intended for random access with short data packages inthe scenario of massive machine-type communications(mMTC).Compared w

169、ith the far-field transmission model applied in multiple access technologies such asSDMA,the near-field transmission model reveals the two-dimensional beam focusing characteristics inthe angle-distance domain and yields a larger spatial degree of freedom,indicating a great increase inavailable spati

170、al resources.Therefore,integrating near-field characteristics into the design ofmulti-access technologies would be more conducive to serving the access requirements of massiveusers and further improving the system spectrum efficiency.Fig.2.14 Schematic diagram of near-field multiple access2.8 Massiv

171、e IoT CommunicationsMassive Internet of things(IoT)communications refer to the networking infrastructure andprotocols required to support the connectivity needs of a large number of IoT devices.These devicestypically generate and transmit small amounts of data intermittently,often with low powercons

172、umption requirements.In such a scenario,massive control type interactions are the dominantcommunication scenario.These control type communications are mostly with short packet that is evenshorter than the signaling data length,but of a massive number.The communication infrastructure mustbe capable o

173、f supporting a massive number of IoT devices,potentially numbering in the millions oreven billions.This requires scalable network architectures,protocols,and management systems thatcan efficiently handle the increased traffic and device density.On the other hand,some high concurrentcommunications mi

174、ght happen with even faster transmission speed requirements,for instance,whileworking together to deliver a real-time industrial situation awareness to the control center in theindustrial IoT applications.By addressing these requirements,communication infrastructure needs to30/156effectively support

175、 the connectivity needs of massive IoT deployments,enabling a wide range ofapplications and unlocking the full potential of IoT technology.ELAA-based Near-field technology is a potential player for massive IoT communications in 6G,especially for the massive connectivity and high-speed concurrent com

176、munication requirements.Byincorporating the ELAA,the base station can connect more IoT devices within the range of theRayleigh distance.Moreover,the ELAA system working on the near-field range will be an idealsolution to the high-speed concurrent communications.According to a prior study 68,arbitrar

177、y signalto noise ratio(SNR)can be achieved simply by increasing the transceiver number,which yields fastertransmission speed while deployed.2.9 On-chip Wireless CommunicationsOn-chip Wireless Communications refers to the utilization of on-chip antennas or near-fieldcoupling,among other wireless inte

178、rconnect methods,to facilitate data exchange and wirelesscommunication between different modules within or among chips.The transmission distance ofon-chip wireless communication typically falls below 1cm,offering advantages such as low loss,hightransmission rates,and high integration.The application

179、s of on-chip communication are widespread,notably in the Internet of Things(IoT)domain,enabling seamless interconnection among smart chips,devices,and wearables,thereby significantly enhancing communication efficiency and reducing wiringcomplexity.However,this technology faces limitations including

180、increased chip area costs,securityand privacy concerns,and increased power consumption.Nevertheless,as the operating frequency risesto the millimeter-wave/terahertz frequency bands,the size of on-chip antennas significantly reduces,leading to a substantial decrease in chip area 6970 Moreover,the pro

181、ximity between transceiverchips reduces,thereby relaxing the signal power requirements for wireless communication.Additionally,the high frequency signals exhibit good directionality,enhancing the security andreliability of information transmission.These factors substantially alleviate the design com

182、plexity ofon-chip wireless communication systems71.Simultaneously,compared to traditional wired communication method between chips,utilizingwireless communication can avoid the drawbacks introduced by transmission lines,such as highlatency,high crosstalk,limited bandwidth,and parasitic effects.As il

183、lustrated in Fig.2.15,byintegrating on-chip antennas into the chip,the signal transmission mode shifts from traditional wiredtransmission to on-chip and inter-chip wireless communication,forming a flexible architecture ofon-chip networks,thereby avoiding the limitations of wired interconnection layo

184、uts72.Fig.2.15 Utilizing on-chip and inter-chip communication with antennasAdditionally,on-chip wireless communication plays a significant role in signal transmissionbetween different system-level chiplets.As illustrated in Fig.2.16,when forming a System-on-a-Chip(SoC)through heterogeneous integrati

185、on,the difficulty of high-frequency interconnection between31/156chiplets arises due to different structures and semiconductor materials.Traditional wire bondingmethods severely degrade signal integrity.By utilizing on-chip wireless communication,high-speed,high-bandwidth heterogeneous integration b

186、etween chiplets can be achieved,effectively enhancing theversatility of large-scale heterogeneous systems73.Fig.2.16 Wireless interconnection between chips with different semiconductor materialsFurthermore,as on-chip wireless communication exploits the radiation effects of on-chip antennasor near-fi

187、eld coupling,it is no longer constrained by the one-to-one data transmission mode oftraditional wired interconnection.In terms of data transmission,it offers higher flexibility andadaptability,enabling one-to-many transmission to support high-density device connections and dataexchanges among massiv

188、e devices,thus providing more feasibility for the design of single-chipmulti-core processors,as illustrated in Fig.2.17 74.Fig.2.17 Single-chip multi-core processors utilizing on-chip wireless communication technologyIn conclusion,on-chip wireless communication can be widely applied in various mobil

189、e devicesand embedded systems,such as smartphones,tablets,smartwatches,IoT devices,etc.,to enhance theperformance of communication devices.The application of on-chip wireless communication in 6G canreduce communication latency,provide faster data transmission capabilities,and enhance the real-timepe

190、rformance and energy efficiency of communication systems.High-speed and efficient on-chipcommunication provide feasible solutions for future 6G wireless communication systems,terahertzintegrated circuits,and chip-to-chip communication interconnections.32/1563 3 Fundamental Theories of Near-fieldWith

191、 the technological evolution from 5G to 6G communication,in order to further improvebeamforming performance and communication rate,larger antenna array apertures and highercommunication frequencies are being employed.However,larger arrays also bring many traditionallyfar-field communication scenario

192、s into the electromagnetic-defined near-field communication range.Innear-field communication,the electromagnetic waves used for information transmission can only beregarded as spherical waves instead of plane waves.This new physical characteristic is inevitable andintroduces many new electromagnetic

193、 effects,such as spatial non-stationarity,polarization,andevanescent waves.As a result,many traditional communication algorithms that were explicitlydesigned for far-field operation suffer from severe performance degradation or fail to leverage the newfeatures for optimal performance in 6G near-fiel

194、d scenarios.In this chapter,starting from the definitionof near-field provided by electromagnetic theory,we analyze the near-field electromagnetic effects,explaining their sources and impacts on existing systems.Furthermore,based on the existing literatureon near-field communication,we summarize the

195、 changes in communication system design andperformance caused by the emergence of near-field effects,focusing primarily on communicationdegrees of freedom and communication capacity.The basic theory of near-field includes four main parts:electromagnetic near-field definition,near-fieldelectromagneti

196、cpropertiesandphysicaleffects,theoreticalanalysisofnear-fieldcommunication degrees of freedom and near-field performance analysis.3.1 Near-field Range PartitioningIn this section,we first introduce the differences between far-field and near-field communications.Then,we establish principles for deter

197、mining the boundaries of far-field and near-field regions inseveral typical application scenarios.As shown inFig.3.1,based on electromagnetic theory and antenna theory,the fields around atransmitter can be divided into the near-field and far-field regions,with the near-field further classifiedinto t

198、he reactive near-field region and the radiative near-field region75.The reactive near-fieldregion is limited to the space close to the antenna(within the Fresnel distance),where evanescentwaves dominate,and the electromagnetic field does not propagate from the antenna in the form ofradiative waves.T

199、he radiative near-field region extends several wavelengths away from the antenna(between the Fresnel distance and the Rayleigh distance).The Rayleigh distance is also known asFraunhofer distance.In this region,the amplitude differences between electromagnetic waves ondifferent antennas within the ar

200、ray are not significant,but the phase changes exponentially with theindex of the antennas.The signal propagation model in this region must be modeled using a sphericalwave model.The far-field region surrounds the radiative near-field region,and in the far-field,electromagnetic waves can be approxima

201、ted as plane waves.Since the reactive near-field region isusually small and evanescent waves decay exponentiallywith distance,practical near-fieldcommunication systems primarily focus on wireless communication within the radiative near-fieldregion,where near-field generally refers to the radiative n

202、ear-field region.33/156In existing research,there are multiple perspectives and empirical rules to characterize theboundaries between the near-field and far-field regions,mainly including phase difference,powerdifference,channel capacity,and localization error.(1)Phase difference perspectiveFrom the

203、 perspective of phase difference,the classic boundary between near-field and far-field isreferred to as the Fraunhofer distance or Rayleigh distance 76(considering a maximum phasedifference between spherical wave and plane wave model not exceeding8),expressed as22,where represents the maximum apertu

204、re of the antenna,and represents the wavelength of the carrier.If thedistance between the user and the base station is greater than the Rayleigh distance,the user can beconsidered to be in the far-field region.In this region,the signal propagation can be approximated asplane waves.On the other hand,

205、if the distance between the user and the base station is smaller than theRayleigh distance,the user can be considered to be in the near-field region.Fig.3.1Far-field plane wavefront and near-field spherical wavefront and corresponding physicalspace normalized received energyPlane waves differ from s

206、pherical waves in their ability to adjust to radiant energy from space.More precisely,plane waves are far-field approximations of spherical waves.In the far-field region,thephase of an electromagnetic wave can be approximated by a Taylor expansion in terms of a linearfunction of the antenna exponent

207、.This clean linear phase forms a plane wavefront that is only relatedto the angle of incidence.Thus,using the plane wavefront,far-field beamforming can steer the beamenergy to specific angles at different distances,which is also known as beam steering.Unfortunately,this clean linear phase does not c

208、ompletely reveal information about the spherical wave.In thenear-field region,the phase of a spherical wave should be derived accurately based on the physicalgeometry,which is a nonlinear function of the antenna index.The information on the angle ofincidence and the distance for each path between th

209、e BS and the UE is contained in this nonlinearphase.By utilizing the additional distance information of the spherical wavefront,near-fieldbeamforming can focus the beam energy at a specific location and achieve energy focusing in both theangle and distance domains.Based on this property,near-field b

210、eamforming is also known as beamfocusing.34/156The primary concept behind the derivation of the Rayleigh distance is as follows 75.The truephase of an electromagnetic wave must be calculated based on an accurate spherical wavefront modeland the BS antenna position.In the far-field case,this phase is

211、 usually approximated by a first-orderTaylor expansion based on a planar wavefront model.This approximation results in a phase differencewhich increases with decreasing distance.The distance between the center of the BS array and thecenter of the UE array is defined as the Rayleigh distance when the

212、 maximum phase differencebetween all BS and UE antennas reaches 8.Therefore,if the communication distance is shorter thanthe Rayleigh distance,the maximum phase difference will be greater than 8.In this case,thefar-field approximation becomes inaccurate,so it is necessary to utilize the near-field p

213、ropagationmodel.Basedonthisdefinition,thenear-fieldrangesofsingle-input-multiple-output(SIMO),multiple-input-single-output(MISO),and multiple-input-multiple-output(MIMO)communicationsystems can be obtained.As shown in Fig.3.2,the near-field range for SIMO/MISO scenarios isaccurately determined by th

214、e classical Rayleigh distance,which is proportional to the square of theaperture of the BS array.For the MIMO scenario,both the BS array aperture and the UE array aperturecontribute to the Rayleigh distance since ELAA is employed on both sides of the BS-UE link;i.e.,thenear-field range is proportion

215、al to the square of the sum of the BS array aperture and the UE arrayaperture.For RIS systems,the cascaded BS-RIS-UE channel consists of BS-RIS and RIS-UE links.Therefore,the 8maximum phase difference needs to be calculated by summing the BS-RISdistances and the RIS-UE distances when calculating the

216、 phase difference,and the near-field range ofthe RIS system is determined by the harmonic mean of the BS-RIS distances and the RIS-UE distances,as shown in Fig.2.It can be further seen from Fig.2 that RIS-assisted communication operates in thenear-field region as long as either of these two distance

217、s is shorter than the Rayleigh distance.Therefore,near-field propagation is more likely to occur in RIS systems 32.Fig.3.2Near-field range for typical communication scenarios(2)Power difference perspectiveWhen using the optimal Maximum Ratio Combining(MRC),signals from different antennaelements can

218、be completely aligned in phase,thus eliminating the impact of phase differences onreceived power.However,due to imperfect channel estimation,MRC may struggle to fully neutralizephase differences.Therefore,considering power losses in practical systems,reference 77 modifiedthe traditional Rayleigh dis

219、tance and proposed the effective Rayleigh distance to characterize theboundary of the near-field range.After eliminating the effect of signal phase on received power through MRC,the received poweris only determined by the amplitude response differences of the antenna elements at the receiver.Conside

220、ring the amplitude response differences of different antenna elements on the same transmitter35/156antenna array,the Critical Distance and Uniform Energy Distance is proposed 7879,whichcharacterize the near-field range from the perspective of power differences between different antennaelements.That

221、is,the power ratio between the weakest and strongest antenna elements detected at thereceiver exceeds a specified threshold beyond this distance.The Critical Distance is determined by theantenna aperture,primarily characterizing the boundary of the field near the antenna aperture axis.TheUniform Ene

222、rgy Distance further considers factors such as array structure and the projected aperture ofthe antenna array,providing a more accurate description of the near-field boundary in off-axis regions.Looking at the differences in received power between the plane wave channel model and thespherical wave c

223、hannel model from another perspective,reference 80 has derived the equal-powerlines and equal-power surface for the near-field region based on the uniform linear array(ULA)and theuniform circular planar array(UCPA)structures respectively,characterizing the near-field range.(3)Capacity perspective:Fr

224、om the perspective of capacity,the near-field range can be described by combining the channelcapacity 81,eigenvalues 82,rank 80,multi-stream transmission characteristics 83,or effectivedegrees of freedom 84,to evaluate the applicable area of far-field plane waves and near-fieldspherical waves.Refere

225、nce 80 proposes the boundary of the near-field region through equip-ranksurface.It shows that the near-field range increases with the number of scatterers in both line-of-sight(LoS)and non-line-of-sight(NLoS)environments,with is more apparent in the NLoS environment.Considering spatial reuse,referen

226、ce 83 introduces the effective reuse distance metric(),representing the maximum distance at which the channel can efficiently accommodate independentspatial streams at a specific signal-to-noise ratio(SNR).The near-field boundary from the perspectiveof multi-stream transmission is discussed by combi

227、ning the channels effective degrees of freedom 84,which demonstrates that the near-field range is not only related to the antenna array aperture but alsoinfluenced by the number of antenna elements.(4)Localization error perspectiveThe Fraunhofer distance serves merely as a rough estimate to delineat

228、e between the far-field andnear-field regions.However,for localization applications,it is not an appropriate boundary betweennear and far fields,as it does not consider several essential parameters for localization,such as AoA,beam squint,or transmit power.In fact,it has been shown in 85 that this d

229、istance is insufficient tosuggest when the mismatched far-field model can be used in practice,instead of the more accuratenear-field,without significant performance degradation due to the model mismatch.Therefore,a metricbased on the mismatched Cramer-Rao lower bound is proposed,such that the bounda

230、ry between thefar-field and the near-field is the-3 dB contour of the model mismatch positioning error between thetwo regimes.3.2 Near-field Electromagnetic Physical EffectsIn the near-field region,the electromagnetic physical effects of wireless signals will becomenon-negligible.Specifically,electr

231、omagnetic physical effects include tri-polarization effects,energymapping effects,etc.Next,we will start with Maxwells equations to establish electromagneticnear-field channels and reveal these electromagnetic near-field effects.36/1563.2.1Near-field Electromagnetic Signal ModelFig.3.3Near-field ele

232、ctromagnetic radiation system diagramAs shown in Fig.3.3,in the Cartesian coordinate systemOXYZ,we consider that the transmitteris located at pointtttt,x y zpinside the source region3t RRand equipped with a currentdensityt()J p.The transmitter is a common low-cost single-polarized antenna with a nor

233、malizedpolarization direction vectorxyzttttxyz(xyz、are three basis vectors).The coordinates ofeach point on the receiving array arerrr,0 x yp.Further,we consider the scalar electric field defined from the power point of view.Specifically,we exploit the scalar electric field that is a component of th

234、e Poynting vector perpendicular to theobservation region:ss0()()exp j,k rrrEEwhereEnergy mapping effects along the-directe22rionTransmitting energyEn rgy Free space mapping attenuattsrt22tinr,tr,t22(,ionfactor)()=()11=4Zyxzt xt yttrrr MEEpprrzppE EGeneral polarization lo2ss ，(3.1)where(,),(,)z xz yM

235、,r,txx,r,tyy,tzz,r,trtxxx,r,trtyyy.0inin2IEis the initial electric intensity measured in volts.In particular,when the transmitter is polarized toward the positiveYaxis,i.e.,ty.(3.1)canbe simplified to37/156r,t222t2ts,Yin221().4=xzzrrrrEE(3.2)When the signal is incident perpendicularly to the receivi

236、ng surface,i.e.,tt0 xyandt1=zr.(3.2)can be simplified tor,t222t2s,Y,vin221().4=xzrrrEE(3.3)Whentryy,i.e.,r,t22t21=xzr,there is no polarization loss,and(3.3)is simplified to22s,Y,vin21().4=rrEE(3.4)(3.4)is the classic Friis formula.Next,we give the classic far-field signal formula:rtrtin0popopo()exp

237、j,2farx xy ykrrrrEE(3.5)where222pottt+=+rxyz,and the phase term uses a second-order Taylor expansion.Further,(3.5)can be simplified toin0 popo()exp j2fark rrr。EE(3.6)From the near-field signal models(3.1)-(3.4)and far-field signal models(3.5)and(3.6),we cansee:For a near-field signal model,the ampli

238、tude term will include the polarization loss of theantenna,the energy mapping coefficient,and the point-to-point free space loss factor,andthe phase term will be determined by the exact point-to-point distance.The far-field signal model is an approximation of the near-field signal model.For thefar-f

239、ield signal model,the amplitude term only has a fixed free space loss factor,and thedistance of the phase term is also a fixed distance from the transmitter to the receiverreference point.3.2.2Near-field Electromagnetic EffectsTri-polarization effect arises naturally for near field communication sin

240、ce the solution to theMaxwells equation for a dipole antenna has a rapid-decaying radial component.Considering thetri-polarization effect,in 86 the authors considered both the near-field spherical wave channel and themulti-polarization effect using a vector Greens function,and based on this the mult

241、i-polarized38/156near-field spherical wave channel as well as the proposed polarization and channel oriented dualprecoding are considered,and the considered systematic diagram is shown as follows.Fig.3.4 Near-field multi-polarized spherical waves modelSimulation results demonstrate that the multi-po

242、larization effect in near-field communication cansignificantly increase the system capacity in a specific range.Fig.3.5 Near-field tri-polarized channel capacityThe near-field evanescent wave effect mainly affects the communication degrees of freedom andcapacity of the reactive near-field,which we w

243、ill describe in the next subsection.For the near-fieldbeam splitting effect,in a near-field RIS,a phase shifter-based beamformer can produce a focusedbeam aligned to a specific location,thus providing beam focusing gain.Such beamformers work wellinnarrowbandsystems.However,forbroadbandsystems,duetot

244、heuseofanalmostfrequency-independent phase shifter,spherical beams of different frequencies are focused at differentphysical locations,which is known as the near-field beam splitting effect.This effect can lead to severearray gain loss because beams of different frequencies cannot be aligned with th

245、e target user at aspecific location,which needs to be carefully considered in the design of broadband systems.Although the beam splitting effect makes it more difficult to align the energy of a broadbandsystem to the user,resulting in a degradation of the beam focusing performance,it has a correspon

246、dingbenefit:since the same guiding frequency corresponds to the generation of spatially multiple beams,itis possible to control the angular range of the beam coverage at different frequencies by designing thesystem parameters.With this benefit,very fast CSI acquisition can be realized in the far-fie

247、ld for fastbeam training or beam tracking.The research on this problem in traditional far-field communication ismainly divided into two types of work:the first type of technique hopes to mitigate the array gain loss39/156caused by far-field beam splitting,and introduces time-delay circuits in the be

248、amforming structure tomitigate the far-field beam-splitting effect;the second type of technique realizes the fast acquisition offar-field CSI in a large-scale multiple-input-multiple-output system by controlling the time-delayparameter and the multiple beams.The effect of near-field beam splitting e

249、ffect is shown inFig.3.6,where it can be seen that thereare multiple energy focusing points spatially during near-field broadband communication.Fig.3.6 Near-field beam splitting effect schematicThe near-field beam splitting effect is defined and analyzed in the paper 87,and a time delay(TD)based bea

250、mformer is utilized to overcome this effect.We propose to divide the whole array intosubarrays and then assume that the user is located in the near-field range of the whole array but in thefar-field range of each subarray.On this basis,delay circuits can also be utilized to compensate for thegroup d

251、elay between different subarrays caused by the near-field spherical wavefront.As a result,thebeam over the entire bandwidth can be focused at the desired spatial angle and distance,and thenear-field beam splitting effect is mitigated accordingly.3.2.3Near-field Beam CharacteristicsAfter studying the

252、 near-field characteristics,the properties of the near-field beam can be analysedand summarized in the following three points:near-field distance-domain focusing,distance-domainasymptotic orthogonality,and distance-domain focusing properties of rectangular and circular array.First,for the near-field

253、 distance-domain focusing characteristics,the depth-of-focus(DF)wascalculated in 88 When a transmitter with distance F is focused using matched filtering,the DF is:+10,10(3.8)where dFAis the array Rayleigh distance,and the depth of the beam depends on where the matchedfilter is focused as shown in F

254、ig.3.7.When the focus is less than dFA/10,the depth of the near-fieldbeam assignment is finite.40/156Fig.3.7 Trend of beam gain with distanceNear-field beam focusing concentrates the beam energy at a specific location determined by angleand distance.In order to utilize the additional spatial distanc

255、e domain resources to improve the spectralefficiency,researchers in 89 demonstrated the asymptotic orthogonality of the near-field arrayresponse vectors,and the channel correlation can be expressed as:+(3.9)where =N2d21221r1r?.This means that as the number of array antennas N tends to infinity,also

256、tends to infinity and fneartends to 0.As shown in Fig.3.8,as the number of antennas increases,the correlation between two array response vectors at different distances at the same angle tends to 0.Fig.3.8 Channel correlation versus antenna curveThe above two beam characteristics are for the Uniform

257、planar array(UPA)and Uniform lineararray(ULA)scenario,respectively.Moreover,the distance-domain focusing characteristics areelaborated next for the uniform circular array(UCA)90.The beam focusing gain in the UCAscenario is approximated as a zeroth-order first-type Bessel function.From Fig.3.9,it can

258、 be seen thatthe ULA beam focusing gain decreases smoothly with decreasing distance,while the beam focusinggain of UCA decreases faster,which indicates that UCA is able to focus the signal power in a smallerrange and mitigate the power leakage.41/156Fig.3.9 Comparison of beamforming gain between UCA

259、 and ULA3.3 Near-field Degree-of-freedom Theoretical AnalysisBased on the Fourier plane wave unfolding channel modelling,the authors in 91 conducteddetailed research on the wave number domain modelling of near-field large-scale antenna arrays,especially investigating the degree of freedom and the co

260、mmunication capacity gain that can bebrought about by the evanescent wave effect in the near-field communication,of which the degree offreedom gain is shown as follows,and there can be a gain of 30%in the typical reactive near-fieldregion.In the left side figure of Fig.3.10,the white wave number poi

261、nts correspond to the plane wavesavailable for far-field communication,the green wave number points correspond to the additionalevanescent wave numbers available for near-field communication,and the red points are theevanescent wave numbers that are not available because of too much attenuation.Sinc

262、e 6G systemswill mostly operate in the radiative near-field,we cannot expect to get more wave number points(spatial degrees of freedom),but we can still make use of them for beamforming in both angle anddepth,which was previously not possible.Fig.3.10 Extra degrees of freedom in the reactive near-fi

263、eldReference 92 proposed a signal space approach to study the number of degrees of freedom of theelectromagnetic field under arbitrary scattering conditions from the point of view of Nyquist sampling.It considers antenna elements in space as spatial sampling points and uses the sampling points requi

264、red42/156to recover the electromagnetic field as communication degrees of freedom.Under isotropic propagationconditions,sampling per square meter is reduced by 13%compared to classical half-wavelengthsampling.This gap increases as the angular selectivity of the scattering increases,resulting in asig

265、nificant reduction in spatial sampling complexity.Fig.3.11 Nyquist sampling under isotropic scattering conditionsIn the context of spatially discrete(SPD)-MIMO,the comprehensive channel response in thenarrowband case is represented as a matrix H.In this case,the number of spatial degrees of freedom(

266、DoFs)of the channel is determined by the number of positive singular values of H or the rank of thecorrelation matrix HHH.In a far-field MIMO LoS channel,the presence of only a single incident angleat all points across the array corresponds to plane-wave propagation.In such cases,the channel has ara

267、nk of 1,corresponding to only a single DoF.In contrast,within the near-field region,spherical wavesmanifest non-linearly varying phase shifts and power levels for each link.This inherent diversityincreases the rank of the channel matrix.This implies that,by reducing the antenna spacing within afixed

268、 aperture,the number of spatial DoFs can be significantly increased 93.However,it is essential to note that when two antennas are in close proximity to each other,thewaves they generate at the receiver become nearly indistinguishable.This limitation should beconsidered,as it could potentially restri

269、ct the achievable increase in channel capacity whenincorporating a large number of antennas within a fixed aperture.For a given aperture,the singularvalues of matrix H presents a two-slope property.We denote the ordered positive singular values ofmatrix H as 1 2 N.Extensive simulations and measureme

270、nts have consistently shown that,for small values of n,the nvalues exhibit a slow decay until they reach a critical threshold,beyondwhich rapid decay occurs.This critical threshold is termed the number of effective DoFs(EDoFs),andis illustrated in Fig.3.12 93.This phenomenon becomes more pronounced

271、as the number oftransceiver antennas increase.By assuming that1 2 e e+1 N 0,weapproximate the number of EDoFs as e tr2HHH/HHHF294.Besides,we note that the number ofEDoFs of near-field channels decreases with the propagation distance.We note that tr2HHH/HHHF2was originally introduced by Verd to evalu

272、ate the slope of the capacity curve in the low bit SNRregime 95.In recent years,some researchers have noticed that this expression can be also used toapproximate the number of EDoFs of near-field channels.Next,we examine the scenario where both transceivers are equipped with continuous-aperture(CAP)

273、arrays,denoted as CAP-MIMO.In contrast to an SPD antenna array,which providesfinite-dimensional signal vectors,the CAP array supports a continuous distribution of source currents43/156within the transmit aperture.In this context,CAP-MIMO can be regarded as a limiting case ofSPD-MIMO with an infinite

274、 number of antennas in a fixed aperture area,but this is achieved throughspatial oversampling,so the number of EDoFs remains the same.For example,the singular values ofthe CAP-MIMO channel exhibit a two-slope trend,as depicted in Fig.3.12 96.Therefore,for thenear-field CAP-MIMO channel,the performan

275、ce is still limited by the number of EDoFs.UnlikeSPD-MIMO,CAP-MIMO channels cannot be characterized by a finite-dimensional matrix.Typically,it is necessary to use Greens functions to characterize the electromagnetic propagation environmentbetween any two points on the transceiver apertures.Obtainin

276、g parallel subchannels requiresperforming eigenvalue decomposition on the kernel function of Greens functions,resulting in highcomputational complexity 96.To estimate the number of EDoFs of near-field CAP-MIMO,one canreplace the channel matrix H in the formula e tr2HHH/HHHF2with Greens functions 94.

277、Additionally,based on existing simulation results,the following conclusion can be drawn:the numberof near-field EDoFs is proportional to the product of the transmitter-receiver aperture area andinversely proportional to the transmission distance between the transmitter and receiver 93.Fig.3.12 Singu

278、lar Values of Near-Field SPD-MIMO.D denotes the transmission distance and N is thenumber of transmit and receive antennas.3.4 Near-field Performance Analysis and Measurement3.4.1Near-field Performance AnalysisTheevolutionfrom5GmassiveMIMOcommunicationto6Gultra-massiveMIMOcommunication involves more

279、than just a simple increase in the number of antennas or array size.Itfundamentally changes the channel characteristics,such as transitioning from the traditional far-fielduniform plane waves to near-field non-uniform spherical waves and from spatial stationarity to spatialnon-stationarity 799798.As

280、 a result,most performance analyses based on the traditional far-fielduniform plane wave model,such as asymptotic channel gains,need to be re-evaluated within thecontext of the new near-field models.Moreover,the authors in 99 investigated a cooperative relayingNFC and FFC coexisting system with XL-a

281、rray,raising an interest question:does the capacityimprovement from NFC compensate the performance loss suffered from more time slots?To providethe solution to this question,99 analyzed the achievable capacity for the proposed NFC schemes44/156compared with FFC,confirming that the capacity improveme

282、nt from NFC compensates theperformance loss from the half-duplex principle.In the traditional far-field model,the equivalent channel gain of a uniform planar array increaseslinearly/quadratically(squared)with the array size until it approaches infinity.This conclusion clearlycontradicts the laws of

283、physics.To obtain more general conclusions,references 799798 andreferences 100101 have proposed new near-field spherical wave propagation models for massiveMIMO active arrays and massive MIMO passive arrays,respectively.These models consider theasymptotic performance when the size of the active/pass

284、ive array tends to infinity.In the sphericalwave model based on near-field communications,the equivalent channel gain exhibits nonlineargrowth with an increasing number of active antennas/passive elements,governed by the new parameterof angular span 102.As the number of active antennas/passive eleme

285、nts tends to infinity,theequivalent channel gain converges to a constant value 7997-101.Compared to co-located massive MIMO arrays,sparse massive antenna arrays have a largerphysical aperture,making their near-field characteristics more prominent.103 studies theperformance of uniform sparse antenna

286、array.By exploiting the non-uniform distribution of the spatialangle difference,it is shown that sparse antenna arrays may achieve better interference suppression andsuper-resolution spatial localization capabilities.Fig.3.13 shows the cumulative distribution functionof communication rates for co-lo

287、cated and uniform sparse antenna arrays,and sparse array can achievemore than four-times data rate than co-located ULA.Fig.3.13 CDF of data rate for co-located and sparse array103Furthermore,104-107 studied a novel modular array architecture to accommodate extremelylarge arrays,termed modular XL-arr

288、ay.All array elements in modular XL-array are regularly deployedin a modular manner on a common platform.Each module is comprised of a moderate/flexible numberof array antennas with the inter-element spacing typically in order of half-wavelength,while differentmodules are separated by the relatively

289、 large inter-module spacing,so as to enable conformal capabilitywith the deployment environment in practice.104105 proposed the near-field non-uniform sphericalwave(NUSW)model for modular XL-array.Under this model,the authors derived the closed-formexpression for near-field SNR,which reveals the SNR

290、 scaling law and asymptotic performance,as wellas the difference from the conventional far-field uniform plane wave(UPW)model.Fig.3.14 shows45/156that the SNR result under UPW model grows linearly unboundedly,while the SNR result under NUSWmodel approaches to a constant value.According to the charac

291、teristics of modular array architecture,106 proposed sub-array based uniform spherical wave(USW)models under different angles/commonangle,and analysed its near-field beam focusing patterns.It can be seen from Fig.3.15 that comparedto co-located counterpart with the same number of antennas,modular XL

292、-array can significantlyenhance spatial resolution from angular and distance dimensions,while at the cost of more severegrating lobes.To further alleviate the issue of grating lobes,107 proposed a user grouping strategybased on greedy algorithm for the multi-user modular XL-MIMO communication system

293、.As a result,users located within the grating lobes of each other are not allocated to the same time-frequencyresource block(RB),which greatly mitigates inter-user interference(IUI)in multi-user scenarios.Ascan be seen in Fig.3.16,in contrast to co-located counterpart,modular XL-MIMO can significant

294、lyenhance the communication performance.Fig.3.14 SNRs versus the number of array elements for different models104.(a)Near-field beam focusing patterns versus spatial frequency differences.46/156(b)Near-field beam focusing patterns versus distance differences.Fig.3.15 Near-field beam focusing pattern

295、s under different array architectures106107.Fig.3.16 Achievable sum rates for modular and co-located arrays versusthe circular radius of user distribution,under far-field and near-field linear beamforming 107.Compared to large-scale active antenna arrays,the passive RIS is more likely to achieve ext

296、remelylarge-scale configurations in practice due to its appealing advantages,such as low cost and low energyconsumption.In communication systems where an extremely large-scale RIS(XL-RIS)is deployednear the base station(BS)side,as illustrated in Fig.3.17,the distances from users to the RIS and theBS

297、 are nearly the same.Moreover,when the IRS is sufficiently large,the effective propagation pathloss of the reflected link is comparable to that of the direct link.Due to the half-space reflection natureof the RIS,a BS equipped with omnidirectional antennas will have at most half of its transmitted p

298、owerreflected by the RIS.Therefore,a desired transmit diversity gain can be further achieved throughdeliberate design of the reflected link100101.If we consider passive beamforming design with intelligent reflective surfaces,the beamforminggain of massive intelligent reflector arrays in the near-fie

299、ld model no longer strictly follows the47/156traditional square-law growth108.Instead,it converges to a constant value as the number ofreflective units tends to infinity 101 109.Fig.3.17 XL-IRS aided communication systemGenerally,the path loss increases with distance following a minimum loss exponen

300、t of 2(i.e.,free-space loss model).In other words,the path loss exhibits a square decay with respect to the distance.However,reference 101 first unveiled that under near-field conditions,when the size of IRS goesinfinity,the equivalent path loss of the reflected link with passive beamforming only de

301、cays with theabsolute value of the distance,i.e.,the equivalent minimum loss exponent is 1(the simulation resultsin Fig.3.18 verified this conclusion).For multi-path XL-MIMO communications,the spatial correlation is of paramount importance forthe second-order statistical channel characterization.The

302、 far-field uniform plane wave(UPW)basedspatial correlation only depends on the power angular spectrum(PAS),which exhibits spatialwide-sense stationarity(SWSS).By contrast,the near-field non-uniform spherical wave based spatialcorrelation depends on both the scatterers angles and distances,i.e.,power

303、 location spectrum(PLS),and SWSS is no longer valid 110111.Fig.3.18 Channel power gain versus the IRS-user distance48/156In addition to near-field communication,the extremely large-scale multiple-input-multiple-output(XL-MIMO)also provides new opportunities for high-precision sensing due to its ultr

304、a-high spatialresolution.In this context,the near-field sensing exhibits more practical scaling laws compared tofar-field models 112.Moreover,when the number of antennas in the XL-array goes infinity,theCramer-Rao Bounds(CRBs)for angle estimations in XL-MIMO radar no longer decrease without limit,bu

305、t converge to a constant value 113.To investigate the fundamental limit of the proposed systemspositioning accuracy,the authors in 141 obtained the Fisher Information Matrix(FIM)and CRLBwhile considering the antenna radiation pattern.The analysis results indicate that the FisherInformation Matrix gr

306、ows quadratically with the size of the RIS(see Fig.3.19).Fig.3.19 RIS-assisted Localization PerformancePizzo,Marzetta and other scholars have proposed to model the communication channelcorresponding to a near-field massive antenna array in the wave numberdomain in thepapers115-117.The main idea is t

307、o reconstruct the HMIMO channel based on Fourier expansionusing a finite number of sampling points of the channel in the wave number domain,as shown in Fig.3.20.Similar to the Fourier transform between the time and frequency domains,the relationshipbetween the spatial and wave number domains is also

308、 described by the Fourier transform,and thespatial domain channel can be characterized by the Fourier transform of the wave number domainchannel,which is denoted by:,=122,?(3.10)where Hakx,ky,x,ydenotes the wave number domain channel,ark,rdenotes the receivedwave vector,as,sdenotes the transmitted w

309、ave vector and h r,sis the spatial domain channel.From the above equation,it can be seen that the channel model consists of three main components,i.e.,the transmit and receive wave vectors and the wave number domain channel.Therefore,the modelingof the spatial domain channel can be equated to the al

310、ternative modeling of the wave number domainchannel,which is given by the following equation:,=12,(3.11)where the wave number domain channel can be expressed in terms of the channel spectral densityS kx,ky,x,yrelated to the scattering environment and antenna arrangement.W kx,ky,x,yrelates to the sto

311、chastic properties of the channel.Wave number domain channels generally have asparse structure,i.e.,they are dominated by a finite number of nonzero coefficients.Based on sampling49/156theory,the wavelength domain channel can be approximated by uniformly sampling a finite integralregion.The accuracy

312、 of the channel approximation depends on the number of points of the regionbeing sampled.As the computational complexity increases,one can obtain a more accuraterepresentation of the channel by generating more samples.Fig.3.20 Fourier plane wave expansion channel modelThe simulation results correspo

313、nding to the spherical wave channel modelling are shown in Fig.3.21,and it can be seen that the far-field Rayleigh fading model is no longer applicable at this point,while the modelled channel model coincides with the physical Clarke model.Fig.3.21 Fourier plane wave expansion channel capacity simul

314、ationThe contribution 118 characterized the theoretical capacity limits of near-field communicationsbased on electromagnetic propagation channels in Fig.3.22.From the Maxwells equations and theHelmholtz wave equation which describe the electromagnetic wave propagation properties,118 usedthe Greens f

315、unction to establish an electromagnetic near-field channel model for extremely large-scalediscrete array with single-polarized antennas.Then,for the single-user scenario,the authors derived theclosed-form capacity limit when the array has an extremely large aperture.They also revealed theimpact of a

316、ntenna polarization mismatch and discrete aperture on the system performance in the nearfield.Additionally,based on the proposed channel model,the authors proposed a more generalexpression for near-field Rayleigh distance,depicting the influence of signal incident angles and50/156non-stationary arra

317、y power on the field boundaries.Furthermore,for the multi-user scenario,theauthors exploited the non-stationary features extracted from the single-user scenario to propose twolow-complexity linear precoding schemes based on the concept of visibility regions and utilizing themethod from graph theory.

318、These algorithms effectively addressed the high computational complexitychallenges associated with extremely large-scale antenna arrays.Fig.3.22 Near-field electromagnetic channel capacity limitThe reference 119 provides a comprehensive summary overview of the current principles andtechnologies of H

319、MIMO array near-field communication at the software and hardware levels,which isconducive to a full understanding of the principles,technological evolution,and development directionof HMIMO near-field communication.Fig.3.23 HMIMO Communication Application Scenarios3.4.2Near-FieldMeasurementandNear-F

320、ieldtoFar-FieldTransformationElectromagnetic scattering characteristics refer to the various information contained in the scatteredwaves formed by the radiation of induced currents on the surface of an object when electromagneticwaves are irradiated.Measurement of electromagnetic scattering characte

321、ristics refers to obtaininginformation such as the radar cross section(RCS)and its statistical characteristics,angular scintillation51/156and its statistical characteristics,polarization scattering matrix,and distribution of multiple scatteringcenters of a target through experimental instruments or

322、professional testing equipment 120121.According to different testing distances,it can be classified as far-field measurement,compact-fieldmeasurement,and near-field measurement.Far-field measurement requires a testing distance of R2d2/,where d is the maximum size of thetarget and is the testing wave

323、length.Therefore,a larger space is required.Outdoor far-field testingsites require a large amount of land resources,and the site is also affected by natural meteorologicalenvironments such as precipitation,light,temperature,humidity,and wind speed.Although indoorfar-field measurement avoids interfer

324、ence from the testing environment,it still requires theconstruction of large-scale darkroom buildings and the extensive installation of absorbing materials,which makes its construction and maintenance costs high.Compact-field measurements occupy less space than far-field measurements.However,thecont

325、raction field requires high technical requirements and limited static area,making conductingfull-scale testing for large targets impossible.Meanwhile,its higher cost makes it not an optimalchoice.Indoor near-field testing is a testing method that has developed in recent decades.Near field refers tot

326、esting distances less than the classical far-field conditions(R 0,the powerdecreases.In contrast,in the far-field,the transmitter can only concentrate power towards an angularorientation of the receiver.As illustrated in Fig.4.26(c),when the distance between the transmitter andthe receiver is suffic

327、iently large,the near and far-field models almost coincide,indicating that thefar-field model serves as an approximation of the near-field model.It is noteworthy that for the giventransmitter antenna configuration,the Fraunhofer distance is approximately 200m.However,thisdistance provides a highly c

328、onservative estimate for the reliable employment of far-field models.Thisdiscussion underscores the accuracy of the near-field model in describing wireless propagation andhighlights its utility in leveraging an additional degree of freedom(the distance d)to enhanceperformance in various applications

329、.Notably,in Fig.4.26(a)with d=0.1m,the far-field model tends tooverestimate the received power.Further elaboration and analysis can be found in 173.Fig.4.26 Comparison of near and far-field models for different Tx-Rx distances 173.72/1565 5 Transmission Technologies of Near-fieldDue to the mismatch

330、between near-field propagation models and existing far-field communicationtechnologies,existing far-field technologies will experience significant performance degradation in thenear-field region.This chapter will introduce near-field communication technologies from the aspectsof channel estimation,b

331、eamforming,codebook design,beam training,multiple access technology,non-coherent detection,system architecture and deployment,and standardization.5.1 Near-Field Channel EstimationAccurate channel state information(CSI)is fundamental for designing accurate beamforming in6G communications,which is a k

332、ey factor in achieving ultra-high-speed transmission.The MassiveMIMO 5G technology was designed around the use of model-agnostic channel estimation,using uplinkpilots,uplink-downlink duality,and the use of non-parametric channel estimation methods.Thesemethods can be used for near-field estimation w

333、ithout any modifications.Nevertheless,there are goodreasons for considering parametric estimation methods in 6G:the increased carrier frequency reducesthe available pilot resources per coherence block and increased number of antennas calls for more pilotresources.This issue can be resolved by shifting to parametric methods,where channels are describedwith a number of parameters that is independent