MIMO in WCDMA and LTE STUDENT BOOK LZT 123 9417 R1A LZT 123 9417 R1A Ericsson 2009-1 -
DISCLAIMER This book is a training document and contains simplifications. Therefore, it must not be considered as a specification of the system. The contents of this document are subject to revision without notice due to ongoing progress in methodology, design and manufacturing. Ericsson assumes no legal responsibility for any error or damage resulting from the usage of this document. This document is not intended to replace the technical documentation that was shipped with your system. Always refer to that technical documentation during operation and maintenance. Ericsson 2009 This document was produced by Ericsson. It is used for training purposes only and may not be copied or reproduced in any manner without the express written consent of Ericsson. This Student Book, LZT 123 9417, R1A supports course number LZU 108 7652. - 2 - Ericsson 2009 LZT 123 9417 R1A
Table of Contents Table of Contents 1 MIMO INTRODUCTION...5 OBJECTIVES:...5 INTRODUCTION...7 MIMO BASICS...7 2 RADIO CHANNEL AND ANTENNA BASICS...13 OBJECTIVES:...13 BASIC RADIO CHANNEL PROPERTIES...15 ANTENNA BASICS...26 PRACTICAL IMPLEMENTATION OF ANTENNAS FOR MIMO...29 3 PRECODING AND SPATIAL MULTIPLEXING...35 OBJECTIVES:...35 PRECODING FOR SPATIAL MULTIPLEXING, TX DIVERSITY AND BEAMFORMING...37 SPATIAL MULTIPLEXING...39 4 MIMO IN WCDMA...45 OBJECTIVES:...45 MIMO IN WCDMA...47 5 MIMO IN LTE...57 OBJECTIVES:...57 MIMO IN LTE...59 LZT 123 9417 R1A Ericsson 2009-3 -
Intentionally Blank - 4 - Ericsson 2009 LZT 123 9417 R1A
1 MIMO Introduction 1 MIMO Introduction OBJECTIVES: On completion of this chapter the student will be able to: Describe the basics of MIMO Explain the reason for multi-antenna processing List the different methods of multi-antenna processing Explain the different multi antenna possibilities Explain the general concepts of beamforming, diversity and spatial multiplexing Explain the concepts of MIMO, SIMO, MISO and SISO Figure 1-1 Objectives LZT 123 9417 R1A 2009 Ericsson - 5 -
Intentionally Blank - 6 - Ericsson 2009 LZT 123 9417 R1A
1 MIMO Introduction INTRODUCTION Multiple antenna solutions can be used in order to increase the spectrum efficiency as well as the peak data rates. Different approaches aim for different purposes, e.g. traditional beamforming and transmitter diversity techniques increase the coverage and capacity. Spatial multiplexing, a technique which requires multiple antennas at both transmitter and receiver, increases the peak data rates and spectrum efficiency up to several hundred percent. Advanced antenna solutions Tx & Rx Diversity Multi-layer transmission (Spatial Multiplexing) Beam-forming TX TX Figure 1-2. Multi antenna features. MIMO BASICS The use of multiple antennas relies on two different principles Improving the SNR (Signal to Noise Ratio) Sharing the SNR This means that in scenarios where the SNR is low, improving the SNR is the way to go. This can be achieved by the use of beamforming, tx diversity and/or rx diversity. Beamforming concentrates the transmitted (and/or received) energy in desired direction(s). LZT 123 9417 R1A 2009 Ericsson - 7 -
Tx diversity achieves diversity against the channel fading by transmitting the information at different times and/or from different antenna locations. Open Loop tx diversity does not exploit any channel information (no feedback from receiver) while closed loop tx diversity uses feedback from the UE in order to maximize the performance. When the SNR is high, the data rate and spectrum efficiency can be increased by e.g. increasing the modulation order (e.g. going from 16-QAM to 64-QAM). This gives us more bits/s/hz. However, the improvement in throughput and spectrum efficiency as a function of the SNR is logarithmic. This means that the throughput saturates at high SNRs, resulting in an excessive need for power/link budget in order to reach high data rates. In that case, it is better to share the SNR (or energy) between different dimensions (layers). This is achieved by spatial multiplexing, where several data streams (layers) are transmitted on the same physical radio resources (simultaneously at the same frequency and code). The streams are separated in the spatial domain instead. Single layer Two layers N+I N+I Power: P SINR: ρ Power: P/2 N+I SINR: ρ/2 Power: P/2 SINR: ρ/2 High SINR satures single layer throughput SINR shared among layers Ideally, multiple layers better (C 2layers > C 1layer ) In practice, MIMO may be worse due to Inter stream cross-talk => lowers MIMO SINR:s SINR imbalance Capacity (bps) C 2layers C 1layer ρ/2 ρ/2 SINR High SINR => greater chance multiple layers is better! Figure 1-3. Single layer vs. Multiple layers. - 8 - Ericsson 2009 LZT 123 9417 R1A
1 MIMO Introduction The data rate only increases logarithmically as a function of the SNR (or SINR Signal to Interference and Noise Ratio), at a high SNR. This is according to the Shannon theorem r data = BW x log 2 (1+SNR) (max data rate r data is equal to the bandwidth, BW, multiplied by the base-2 logarithm of the SNR plus 1). On the other hand, at a low SNR, the max data rate increases almost linearly. Therefore, it is not efficient aiming only to obtain a high SNR. It is more efficient to try to create several data pipes with lower SNR (sharing SNR), which will lead to a multiplication of the maximum achievable data rate with up to the channel rank r max. Without MIMO Single stream Rx/Tx processing, Multiple Rx and/or Tx antennas With MIMO Multistream Rx/Tx processing, Multiple Rx and Tx antennas C log(1+snr) C log(1+n x SNR) C mimo N x log(1+snr) Cap Cap SNR N x SNR SNR SNR N x SNR SNR Low SNR: approx linear increase in rate High SNR: logarithmic increase in rate Figure 1-4. Why MIMO? Share SNR between streams Linear increase in rate A traditional way of sharing the SNR is actually by spread spectrum techniques, e.g. CDMA, where the transmission is multiplexed over a wider bandwidth. Different configurations of multiple antennas are shown in Figure 1-5. These include SISO (Single Input Single Output), MISO (Multiple Input Single Output), SIMO (Single Input Single Output) and of course, MIMO (Multiple Input Multiple Output). The naming convention refers to input/output of the radio channel. This means that the transmitter antenna(s) correspond to input to the radio channel and the receiver antenna(s) reception correspond to the output of the radio channel. LZT 123 9417 R1A 2009 Ericsson - 9 -
TX RX S-P RX SISO MISO TX RX S-P RX SIMO MIMO Figure 1-5. Antenna configurations. With multiple antennas at the transmitter and only single antenna at the receiver (referred to as MISO) it is possible to obtain so called beamforming. With this method the transmission signal is steered in a beneficial direction (typically towards the UE). This is accomplished by adjusting the phase (and sometimes amplitude) of the different antenna elements by multiplying the signal with complex weights. This method increases the SNR (Signal to Noise Ratio) and thus the capacity. With this configuration it is also possible to achieve Transmit Diversity. This is done by transmitting time-shifted copies of the signal and thus achieving diversity in the time-domain. This method also increases the SNR. With multiple antennas at the receiver (SIMO or MIMO), it is possible to use receive diversity. A combining method (typically MRC Maximum Ratio Combining) is applied to increase the SNR of the received signal. With multiple antennas at both transmitter and receiver, it is possible to use all of the above mentioned methods. However, with multiple antennas at both transmitter and receiver, it is also possible to achieve spatial multiplexing, also referred to as MIMO. This method creates several layers, or data pipes in the radio interface. The maximum number of layers that can be created depends on the radio channel characteristics and the number of tx and rx antennas. The maximum number of layers that the radio channel can support is equal to the channel rank. The maximum number of layers that effectively can be used is equal or less than the minimum number of antenna elements at the tx or rx side or the channel rank. The number of layers that actually is used for transmission is referred to as the transmission rank. - 10 - Ericsson 2009 LZT 123 9417 R1A
1 MIMO Introduction The data rate can at optimal circumstances be multiplied by the number of layers. Directivity Antenna/Beamforming gain Example Diversity Reduce fading Example Spatial Multiplexing Data Rate multiplication Example Delay S-P Channel knowledge (average/instant) Transmit the signal in the best direction Transmit the signal in all directions Transmit several signals in different directions Different techniques make different assumptions on channel knowledge at rx and tx Many technqiues can realize several benefits Realized benefit depends on channel (incl. antenna) and interference properties Figure 1-6. Multi-antenna possibilities. Better data rate coverage and capacity Directivity and diversity improves link budget Potential for higher data rates Spatial domain provides extra dimension Extra dimension offers increased data rates Ideally: r max times higher data rate than single Tx (r max = min{#tx antennas, #rx antennas}) Transmit r r max parallel symbol streams (layers) Max r max parallel data pipes Number of layers r depends on channel properties Channel rank r: #effective data pipes channel can support (data pipes with sufficiently strong SINR) Rank adaption dynamically adjusts #layers Higher spectral efficiency! Up to r max times higher throughput! Figure 1-7. Possible multi-antenna benefits. In a simple beamforming example, the complex weights are adjusted in order to align the carrier phase in such a way that the individual antenna elements signals are added constructively at the receiver side. With two antenna elements we need two complex weights, here denoted w 1 and w 2. These weights are adjusted with the support of some kind of feedback from the receiver (in this case the UE), in order to maximize the reception. LZT 123 9417 R1A 2009 Ericsson - 11 -
With multiple antennas (here called antenna ports) at both tx and rx side and multiple layers, the complex weights will form a matrix. In case of 4x4 MIMO and four layers, a 4x4 matrix will be needed. This means that each layer will have its own complex weight vector of length equal to the number of antenna elements. Adapt spatial properties of transmission to match current channel conditions Precoding/beamforming example h 1 s w 1 h 2 y = (h 1 w 1 + h 2 w 2 )s + e w 2 Coherent addition of signals at receive side increases SNR Similar to closed-loop TxD in WCDMA Precoding with multi layers Multiply each layer with corresponding precoding vector => matrix weighting! Figure 1-8. Precoding for exploiting channel info at tx side. - 12 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics 2 Radio Channel and Antenna basics OBJECTIVES: On completion of this chapter the student will be able to: Describe the radio channel and antenna basics Explain multi-path propagation Explain time dispersion and delay spread Explain the doppler effect Explain coherence bandwidth and coherence time Explain DOA (Direction of Arrival) DOD (Direction of Departure) and angular spread Explain polarization properties of the radio channel Explain basic antenna properties Explain polarization properties of antennas Describe beamforming using an ULA (Uniform Linear Array) Explain polarization diversity Figure 2-1. Objectives. LZT 123 9417 R1A 2009 Ericsson - 13 -
Intentionally Blank - 14 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics BASIC RADIO CHANNEL PROPERTIES In order to understand the principles of MIMO and spatial division multiplexing (SDM) and spatial division multiple access (SDMA), the radio channel properties is first explained below. The radio channel distorts the transmitted signal between the transmitter and the receiver. This distortion is due to various reasons. For example, the original signal is added with noise and interference, distorted by Doppler shift, time dispersion and frequency selective fading. Noise and interference Reduced coverage and quality Fading (fast and slow) Fluctuating signal strength Time dispersion Inter symbol interference (ISI) and fast fading Doppler shift Frequency errors and reduced performance Radio channel Interference Fast Fading Time dispersion Noise Transmission Slow Fading Doppler shift Figure 2-2. Radio channel impairments. Noise and interference Noise can be divided into internal noise and external noise. Internal noise originates from the system itself, especially from the receiver circuitry. This noise is to a large extent coming from the first amplifier in the receiver. The quality of this amplifier judges the amount of thermal noise added from the amplifier itself. The lower noise factor (NF) the amplifier has, the less noise will be added. This is the reason why we call it LNA (Low Noise Amplifier). External noise comes from other systems, the surrounding environment, galactic noise etc. LZT 123 9417 R1A 2009 Ericsson - 15 -
Interference is typically referred to as interfering signals coming from the own system (e.g. neighboring or own cells) or other systems. External noise Internal noise Own cell interference Neighbor cell interference Transmitted Signal Received Signal + Noise + Interference Figure 2-3. Noise and interference. Quality measures of the wanted signal include for example SNR (Signal to Noise Ratio), SIR (Signal to Interference Ratio) or SINR (Signal to Interference and Noise Ratio). These describe the ratio between the power of the wanted signal and the noise and/or interference. For example, a SNR of 20 db means that the desired signal S is 20 db stronger than the noise N. 20 db corresponds to 100 times more power, since 10*log(S/N) = 10*log100 = 20 db. SNR and SINR are illustrated in Figure 2-4. S P1 + P2 + P3 SIR = = I I S P1 + P2 + P3 SINR = = I + N I + N Transmitted Signal Received Signal + Noise + Interference Power Power P 1 P 2 P 3 f N+I Noise + Interference f Figure 2-4. SIR and SINR A radio channel that is only distorted by white Gaussian noise is referred to as an AWGN (Additive White Gaussian Noise) channel. - 16 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics Multipath propagation When the signal bounces on different obstacles in the environment, different time delayed copies with different powers will reach the receiver. This is referred to as multipath propagation. Multipath propagation gives rise to Fading (fast fading, Rayleigh fading) Time dispersion (Inter Symbol Interference, ISI) Many radio propagation effects such as reflections can attenuate the transmitted radio signal Figure 2-5. Figure 2-5: Multipath Propagation. The received signal contains many time-delayed replicas. Multipath Propagation gives rise to: 1. InterSymbol Interference (ISI) 2. Fast fading (Rayleigh fading) Power (db) Impulse response P0 P1 P2 P 0,τ 0 P 1,τ 1 P 2,τ 2 τ0 τ1 τ2 t(μs) Figure 2-6. Multipath propagation and the resulting impulse response. LZT 123 9417 R1A 2009 Ericsson - 17 -
This occurs when the propagation wave reflects on an object, which is large compared to the wavelength, for example, the surface of the earth, buildings, walls, etc. This phenomenon is called multipath propagation and it has several effects, these are: Rapid changes in signal strength over a small area or time interval Random frequency modulation due to varying Doppler shifts on different multipath signals. Time dispersion caused by multipath propagation delays Previous symbol leaks into current symbol due to the different path delays When the amount of ISI exceeds a certain level (~10%) bit errors occur Can be reduced with equalizers, rake receivers or the use of OFDM ISI Direct signal Reflected signal Path delay difference Figure 2-7 Inter Symbol Interference (ISI). Multipath propagation yields signal paths of different lengths with different times of arrival at the receiver. Typical values of time delays (µs) are 0.2 in Open environment, 0.5 Suburban and 3 in Urban. Direct Signal Reflected Signal Combined Signal Figure 2-8: Multipath Fading. - 18 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics The combination of direct and out-of-phase reflected waves at the receiver yields attenuated signals (Figure 2-8). This results in a varying received signal power as illustrated in Figure 2-9. Power Time between fades is related to RF frequency Geometry of multipath vectors Vehicle speed: Up to 4 fades/sec per kilometer/hour path loss Rayleigh Deep fade caused by destructive summation of two or more multipath reflections distance Figure 2-9 Path loss and fast fading. This attenuation can result in bit errors that occur in consecutive blocks of data (burst errors). As a result the decoder fails to recover such errors. The radio channel characteristics in the frequency domain depend on the time difference of the different components of the received signal. A measurement of the average time dispersion of the radio channel can be defined by the rms delay spread, σ rms. The delay spread is a weighted average of the power and relative delay of the multipath components, as described by the following equation: σ rms = 2 τ P P τ τ τp P τ τ, where P τ is the power at delay τ. The excess delay is typicaslly referred to as the maximum time difference between different multipath components above a certain threshold. LZT 123 9417 R1A 2009 Ericsson - 19 -
On average, the coherence bandwidth is equal to the inverse of the delay spread. B C = 1 σ rms The coherence bandwidth defines how frequency selective the radio channel is. For example, a coherence bandwidth higher than the transmitted bandwidth (which typically occurs when the delay spread is shorter than the symbol time) will give almost flat fading. Flat fading means that the whole transmitted bandwidth fades equally, or in other words, the correlation between all frequencies in the transmitted bandwidth is high. Coherence BW > Tx BW: The transmission experiences almost flat fading Only frequency hopping and/or Antenna diversity (Tx & Rx diversity, beamforming, SM) can reduce and even exploit the fading P Large Coherence BW Coherence BW < Tx BW: FEC can reduce the errors caused by fading Frequency hopping and/or Antenna diversity (Tx & Rx diversity, beamforming, SM) can further reduce and exploit the fading P Tx BW Low Coherence BW f Tx BW f Figure 2-10. Coherence bandwidth. Tx (and Rx) diversity and beamforming can reduce the fading by exploiting the differences between the antenna elements. MIMO with spatial multiplexing can reduce and even exploit the fading by amplifying wanted paths and cancelling unwanted paths in the spatial domain by altering the antenna phases. - 20 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics The time varying radio channel When the terminal is moving, the radio channel properties will change locally. However, the long term statistics of the channel will typically remain the same on average, e.g. delay spread, coherence bandwidth etc. This is referred to as a WSSUS (Wide Sense Stationary Uncorrelated Scattering) channel. As the terminal moves, the fast fading has a large impact on the received signal strength. Locally, the signal strength may vary as much as 40 db (10000 times) due to Rayleigh fading. Rayleigh fading typically occurs when a large number (>10) of uncorrelated signal components are added. This is actually the normal case in urban environments. Another phenomenon that occurs due to mobility is Doppler shifts. When the terminal is moving towards the transmitter/receiver (base-station), the carrier frequency is slightly increased and viceversa. The Doppler shift is equal to the speed divided by the wavelength of the carrier signal. This ratio is valid when the terminal is moving straight towards the transmitter. When moving in a certain angle α towards the transmitter, the following equation is valid. f d = v λ cos ( α ) The inverse of the average Doppler spread is equal to the coherence time of the radio channel. The coherence time describes how slowly the radio channel changes. T C = 1 f d λ = v cos ( α ) A high coherence time means that the radio channel does not change rapidly. The symbol time should typically be shorter than the coherence time. Otherwise the symbols become distorted. In order for MIMO in e.g. LTE to work well, the symbol time should be long enough in order to experience flat fading, but should at the same time not exceed the coherence time of the channel. Also, the feedback from the terminal must be quick and fresh enough so that the radio channel properties have not changed too much when the signal is transmitted. LZT 123 9417 R1A 2009 Ericsson - 21 -
The spatio-temporal radio channel An actual radio channel has certain properties in the spatial domain as well as in the temporal (time) domain. In the spatial domain, the channel can be described in terms of e.g. angular spread of DOA (Direction Of Arrival) and DOD (Direction Of Departure). The angular spread is a weighted average of the DOA or DOD of the radio channel. Typically, the angular spread can be quite low at the base-station (especially if it is mounted on a roof top). The UE will typically have a larger angular spread, since the reflections often comes from objects surrounding the UE in almost all directions. Angular spread at base station side typically much lower than at the terminal side Beamforming at base station side beneficial Easier to obtain low spatial correlation at terminal side (shorter antenna distance than at base station) Angular spread low Angular spread high Figure 2-11. Angular spread in a typical urban scenario. The angular spread AS can be calculated in a similar manner as the delay spread, i.e.: AS rms = 2 ϕ P P τ τ ϕp P τ τ, where φ is the azimuth angle of each multipath component. With a high angular spread, the antenna elements become more uncorrelated, given that they are separated in the same dimension as the angular spread. For example, if the azimuth angular spread is high, then the antenna elements become quite un-correlated even if the separation in the horizontal plane is quite small. In the temporal domain, the channel can be described in terms of e.g. delay spread, excess delay and coherence time. - 22 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics Polarization The radio signal can be transmitted with different polarizations. The polarization is defined by the direction of the electrical field vector (E-field). Orthogonal polarizations, such as e.g. horizontal/vertical or ±45, makes it possible to transmit and receive two different data streams independent of each other on the same frequency at the same time. However, the radio channel usually alters the polarizations so they will interfere with each other. Also, a typical cross polarized antenna will not separate the different polarizations perfectly. The separation depends on the antennas XPD (Cross Polarization Discrimination). A typical value of XPD is approximately 20 db (100 times in power), Vertical transmitter antenna polarization: E Receiver antenna polarization: E Vertical: Full reception The electrical field (E-field) determines the polarization Horizontal: No reception E Figure 2-12. Example of linear vertical polarization. As opposite to linear polarization, circular and elliptical polarization also exists. With circular polarization, the E-field rotates in a circle (constant amplitude), either clockwise (right hand circular polarization) or anti-clockwise (left hand circular polarization), when viewed along the propagation path as seen from the transmitter. With elliptical polarization, the vector describes an ellipsoid. LZT 123 9417 R1A 2009 Ericsson - 23 -
Can be created by two perpendicular linearly polarized antenna elements fed with different phases Figure 2-13. Circular polarization. Figure 2-14 gives a more detailed illustration of circular polarization. Right-hand circular polarized current source Two orthogonal dipoles in phase quadrature Horizontal and vertical Figure 2-14. Antenna polarization. When the plane of polarization is viewed in the direction of propagation at a fixed point in space, if the extremes of the E-field vector describes a circle and rotates clockwise as a function of time, the sense is right-hand circular (RHC). Circular polarization may very well be utilized in cellular systems. The benefit is that the transmission will be more independent of the UE position (the base station has no way of knowing the polarization of the UE antenna(s)). The drawback is that there will be a loss (compared to perfectly co-polarized tx and rx antennas) of 3dB. - 24 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics Reciprocity The radio channel is reciprocal, meaning that the properties of the radio channel are the same, regardless of the direction (UL or DL) of the transmissions. This means that properties like impulse response, delay spread, azimuth spread, fading etc are the same in both UL and DL. This assumes that the same frequency band is used for both UL and DL (as in the TDD mode, for example) and that that the transmissions are performed at the same time. Since this almost never happens, a more loose definition can be used. A radio channel could for example be regarded as more or less reciprocal if the duplex distance is shorter than the coherence bandwidth and/or the time between UL/DL transmissions is less than the coherence time of the radio channel. LZT 123 9417 R1A 2009 Ericsson - 25 -
ANTENNA BASICS The simplest example of an antenna is probably the halfwavelength dipole. It consists of a rod, which is split and fed with the carrier wave. The electrical current in the antenna element creates an electrical and a magnetic field that propagates from the antenna. The electrical field and the magnetic field vectors are perpendicular. An antenna can be seen as a matching of the wave impedances between the antenna feeder (often 50Ω) and the surroundings (free space) of the antenna. The free space impedance is by definition the ratio between the electrical field and the magnetic field absolute values. This wave impedance equals 377 Ω. At a perfect match, no energy is reflected from the antenna back to the transmitter. The half-wavelength dipole has an omni-directional radiation pattern in the azimuth plane. In the elevation plane the radiation pattern is slightly directed perpendicularly from the antenna element. In the direction of the antenna rod the radiation is zero. See figure Figure 2-15. λ/2 E - omni horizontal coverage - wide vertical coverage -low gain Figure 2-15. The classical half wavelength dipole. Omni-directional radiation pattern (Doughnut-shape) The electrical field from the half wavelength dipole is aligned with the direction of the rod (vertical in Figure 2-15). - 26 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics Antenna arrays and beamforming When two or more antenna elements are used for transmission or reception (the antennas are also reciprocal), their respective signals will be summed. The resultant signal vector will depend on their relative position, phase, and the direction of the transmitter/receiver. Also, the polarization will affect this summation. In some directions we will see amplification of the resultant signal and in other directions we will see cancellation (nulls). This is referred to as the array factor. The array factor can be multiplied with the antenna element s radiation pattern in order to obtain the resulting array pattern. If the individual antenna element s radiation pattern is omni-directional (in e.g. the azimuth plane), the resulting diagram will equal the array factor. The most common way of implementing an antenna array for beamforming is by the use of a ULA (Uniform Linear Array), consisting of (cross-polarized) dipoles. The antenna elements (dipoles) will add constructively at certain directions, depending on the geometry of the array and the individual phases of the signals fed to the different elements. When stacking columns of elements horizontally, it is possible to steer the antenna beam in the azimuth (horizontal) plane. The horizontal distance between the elements should be in the order of half a wavelength in order to avoid grating lobes (unwanted sidelobes). Figure shows a 4-element Uniform Linear Array (ULA) When x is λ/2 we have nulls (assuming signals to the different elements in phase) Antenna element distance of d=λ/2 eliminates grating lobes x x Δx = d sin(α) d α d α Figure 2-16. Uniform Linear Array (ULA). LZT 123 9417 R1A 2009 Ericsson - 27 -
When the inter-element distance exceeds half a wavelength, grating lobes are created. With a larger antenna element distance, the main lobe is narrower, but unwanted sidelobes (grating lobes) are created d=0.63λ d=0.5λ d Figure 2-17. Main lobes and grating lobes. When the inter-antenna element distance is approximately one wavelength, the grating lobes become as strong as the main lobes. When increasing the antenna distance further, several lobes with the same strength are created. This may sometimes be beneficial when used with spatial multiplexing. More main lobes are created if the antenna distance exceeds approximately half a wavelength d = 0.5 wavelengths d = 1.0 wavelengths d d Figure 2-18. Two element array radiation pattern. - 28 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics PRACTICAL IMPLEMENTATION OF ANTENNAS FOR MIMO How can we practically implement an antenna configuration for MIMO? Well, a suitable way of achieving uncorrelated antenna elements is to use polarization diversity with cross-polarization. 2x2 MIMO can quite easily be configured with existing crosspolarized antennas. This works well even in line-of-sight (LOS) environments (4x4 MIMO may actually perform worse in a LOS environment than in a non-los environment, since the LOS environment probably lack the useful multi-path diversity and we only have two orthogonal polarizations). A four antenna-port configuration can be realized using two cross-polarized antennas. Many different possible antenna configurations Only two different ones considered here Exploit orthogonal polarizations Reduces inter-layer interference Dual-layer transmission possible also in line-of-sight Line-of-sight not uncommon in urban environments Two cross-polarized antennas Two pairs of cross-polarized antennas Dual stream spatial multiplexing Array gain by means of beamforming on pairs of co-polarized antennas Limited radome size 0.5λ Figure 2-19. Antenna setup at the base station. Figure 2-20 shows a Kathrein cross-polarized, stacked dipole antenna. Kathrein 800 10204 900 MHz Linear array seven dipole pairs ~0.95 λ (wavelengths) apart Figure 2-20. Linear cross polarized array antenna. LZT 123 9417 R1A 2009 Ericsson - 29 -
The cross polarized antenna elements work in pairs, as illustrated in Figure 2-21. Basic building block: Dipole over ground plane Two per polarization Pol. 1: - 45 Pol. 2: + 45 λ/2 Figure 2-21. Cross polarized dipole antennas. MIMO channel properties affecting the choice of transmission scheme The propagation environment obviously affects the properties of the MIMO channel via many factors including mobility, path loss, shadow fading, polarization, angular spread of direction of departures (DOD) and arrivals (DOA). This in combination with the particular antenna setup at the transmitter and receiver determines the channel characteristics. Some of the key parameters regarding an idealized antenna setup are Distances between antennas within the arrays Polarization of the antennas Basically, a larger inter-antenna distance gives less correlated channel for a given fixed angular spread and vice versa. Signals transmitted with horizontal and vertical polarization tend to experience uncorrelated fading and this can also be exploited to reduce spatial correlation if so desired. Another important characteristic of transmitting some signals with horizontal polarization and other with vertical polarization is that often the two orthogonal polarizations remain rather well separated even after reaching the receiver. Under ideal conditions in line-of-sight, they would for example be completely separated. Thus, varying inter-antenna distances and polarization can be used to affect the spatial correlation and the isolation between signals. - 30 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics Since a transmission scheme usually work well in a channel with certain properties and less well with other properties, the antenna setup greatly affects which transmission scheme to choose, and vice verse. The previously mentioned transmission schemes are targeting different channel characteristics and hence are suitable together with different antenna setups. Some of the more obvious possible combinations under idealized assumptions regarding the antennas are listed below Beamforming: o Strong spatial correlation on the transmitter side meaning small inter-antenna distance and co-polarized antennas on transmit side Transmit diversity, e.g. the Alamouti code: o Low spatial correlation on at least the transmitter side achieved by either of co-polarized antennas on both transmit and receive side with large inter-antenna distance on transmit side orthogonally polarized antennas on transmit side and receive side SU-MIMO spatial multiplexing: o Low spatial correlation on both the transmit and receive side and/or good isolation between different transmit-receive antenna pairs provided by either co-polarized antennas on both transmit and receive side with large inter-antenna distances on both sides orthogonally polarized antennas on transmit side and receive side LZT 123 9417 R1A 2009 Ericsson - 31 -
Beamforming: Small inter-antenna distance and co-polarized antennas on tx side Transmit diversity, e.g. the Alamouti code: Co-polarized antennas on both tx and rx side with large inter-antenna distance on tx side Orthogonally polarized antennas on tx side and rx side SU-MIMO spatial multiplexing: Co-polarized antennas on both tx and rx side with large inter-antenna distances on both sides Orthogonally polarized antennas on tx side and rx side Figure 2-22. Antenna configurations Note that the meaning of large and small should be interpreted relative to the angular spread on the intended side of the link. For a base station mounted above roof tops the angular spread might for example be quite small and small may then be taken as half a wavelength and large might be 4-10 wavelengths while on the UE side, which typically experiences a much larger angular spread, half a wavelength might be considered large. Reality is unfortunately not as idealized as indicated above, but the list still gives guidance as to which kind of antenna setups the different schemes prefer. In practice, it is for example difficult to guarantee co-polarized antennas on both transmit and receive side simply because the user might have rotated the device. In the UE, it is often also challenging to design purely co-polarized or purely orthogonally polarized antennas, due to for example form factor requirements and RF interactions. Antenna design is an interesting and important topic that clearly needs further study. There are many ways to set up the antenna arrays. The leftmost configuration in Figure 2-23 shows a suitable way of combining quite good performance for both beamforming and two layer spatial multiplexing. Since the column distance is around one wavelength (theoretically it should be half a wavelength, but the directivity of the individual columns enables the distance to be slightly larger), this configuration has a low grating lobe level. The configuration in the middle of Figure 2-23 may be more suitable for four layer spatial multiplexing and/or tx diversity, since the columns probably will be less correlated. The rightmost configuration of Figure 2-23 may work very well for beamforming and four (or even eight) layer spatial multiplexing, but is on the other hand quite bulky. - 32 - Ericsson 2009 LZT 123 9417 R1A
2 Radio Channel and Antenna basics Dual polarization + Spatial separation (MIMO/diversity) Quad antenna (two columns) Array antenna (four columns) ~λ >>λ 0.7λ Figure 2-23. Sample antenna configurations. Figure 2-24 shows an example of a four column, dual polarized antenna array. Four columns Dual-polarized A few cables... Figure 2-24. Example of antenna array. LZT 123 9417 R1A 2009 Ericsson - 33 -
Intentionally Blank - 34 - Ericsson 2009 LZT 123 9417 R1A
3 Precoding and Spatial Multiplexing 3 Precoding and Spatial Multiplexing OBJECTIVES: On completion of this chapter the student will be able to: Explain the concepts of precoding and spatial multiplexing Explain the concept of spatial multiplexing Explain SDMA (Spatial Division Multiple Access) Explain the difference of single-rank and multi-rank transmissions Explain the concepts of channel rank, transmission rank and layers Describe the difference of antenna ports and antenna elements Explain the role of the precoder Figure 3-1. Objectives. LZT 123 9417 R1A 2009 Ericsson - 35 -
Intentionally Blank - 36 - Ericsson 2009 LZT 123 9417 R1A
3 Precoding and Spatial Multiplexing PRECODING FOR SPATIAL MULTIPLEXING, TX DIVERSITY AND BEAMFORMING The use of a matrix of complex weights is referred to as precoding. The signal is pre-coded with the weight matrix before it is transmitted. The number of rows of the precoder matrix corresponds to the number of antenna ports and the number of columns corresponds to the number of layers. The signals are treated in the frequency domain. In the time domain, this corresponds to convolution. See Figure 3-2. Precoder is a complex matrix, W The rows correspond to the complex antenna weights The columns correspond to layers Layer 1 Layer 2 s 1 s 2 Precoder W w 11 w 21 w12 w 22 a 1 a 2 w W = w 1 2 11 12 a = Ws w21 w 22 a1 w11 w a = 2 w12 w a = w s + w s 11 1 12 1 21 22 21 2 a = w s + w s Weights of antenna port 1 (ɑ 1 ) Weights of antenna port 2 (ɑ 2 ) 22 2 s1 s 2 Figure 3-2. Precoder and matrix algebra. The radio channel with a MIMO configuration can also be described with a matrix. Here the number of rows corresponds to the number of receiver antennas and the number of columns corresponds to the number of transmit antennas. Therefore, the matrix size is N R xn T. See Figure 3-3. LZT 123 9417 R1A 2009 Ericsson - 37 -
y: received signal, a: transmitted signal on antenna port, h: radio channel transfer function a h y = ha With MIMO, the radio channel transfer function becomes a matrix, H a 1 a 2 Radio channel H h 11 h 22 h 12 h 21 Figure 3-3. Radio channel and matrix algebra. y 1 y 2 h H = h 1 2 11 y = Ha 21 11 1 21 1 h12 h 22 y1 h11 h12 a1 y = 2 h21 h 22 a 2 y = h a + h a 12 y = h a + h a The complete transfer function is equal to the product (in the frequency domain) of the information signal s and the precoder matrix and the channel matrix. 22 2 2 y = HWs = h h 11 21 h h 12 22 w w 11 12 w w 21 22 s s 1 2 s 1 Precoder W w 11 w 21 w 12 a 1 Radio channel H h 11 h 12 h21 Receiver y 1 s 2 w 22 a 2 h 22 y 2 Figure 3-4. Matrix operations. In the receiver, the radio channel is estimated and compensated for. This corresponds to an estimation of the inverted channel transfer function and an inversion of the precoder matrix. The inverted radio channel transfer function can be estimated with e.g. a MMSE (Minimum Mean Square Error) equalizer. - 38 - Ericsson 2009 LZT 123 9417 R1A
3 Precoding and Spatial Multiplexing LZT 123 9417 R1A 2009 Ericsson - 39 - y H W s I W W Ws W y H W Ws y H I H H HWs H y H HWs y 1 1 1 1 1 1 1 1 1 1 ˆ ) ( ˆ ˆ ) ˆ ( ˆ ˆ = = = Radio channel is estimated and inversed: Ĥ -1 Inverse precoder matrix applied at receiver: W -1 Figure 3-5. Channel estimation and antenna demapping at receiver side. SPATIAL MULTIPLEXING MIMO (Multiple Input Multiple Output) refer to the multiple antenna configuration or setup. In its simplest case we have two tx antennas and two rx antennas. This is referred to as 2x2 MIMO. In general terms, N T xn R means number of transmitter antennas (intputs to the radio channel) and number of receiver antennas (outputs from the radio channel). Figure 3-6 shows an example of a choice of optimum precoder matrix for transmission rank 2 in case of perfectly cross-polarized antennas and an ideal radio channel. 2 2 1 1 1 0 0 1 1 0 0 1 s y s y s y W = = = = 1 1 0 0 s 1 s 2 y 1 y 2 Figure 3-6. Example of cross polarized spatial multiplexing with rank 2.
The transmission of multiple layers can partly be illustrated by a kind of multi-layer beamforming. The beamforming is performed in the spatial domain. Figure 3-7 shows an example of the antenna radiation patterns for two different precoders (from LTE). The antenna element distance is 0.53 wavelengths. 120 Precoder 1 Precoder 2 90 1 90 1 45 120 45 60 layer 1 layer 2 150 0.5 30 150 0.5 30 180 0 180 0 210 330 210 330 240 270 300 240 270 300 antenna distance = 0.53 Figure 3-7. Antenna diagrams for spatial multiplexing. Antenna distance 0.53 wavelenghts Figure 3-8 shows antenna radiation patterns with the same precoders as the previous figure, but now with antenna element distance of 3.3 wavelengths. We can clearly see that many narrow lobes are formed. This is in traditional cases not wanted, but with spatial multiplexing it can be beneficial. - 40 - Ericsson 2009 LZT 123 9417 R1A
3 Precoding and Spatial Multiplexing Precoder 1 Precoder 2 120 90 1 60 layer 1 layer 2 120 90 1 60 layer 1 layer 2 150 0.5 30 150 0.5 30 180 0 180 0 210 330 210 330 240 300 240 300 270 270 lambda = 3.3 Figure 3-8. Antenna diagrams for spatial multiplexing. Antenna distance 3.3 wavelengths. We can see in both previous figures that the antenna patterns are orthogonal in order to support independent layers. Figure 3-9 shows a simple example of spatial multiplexing where different orthogonal beams transmits and receives two different layers. 120 Layer 1 90 1 60 0.8 150 120 90 1 0.8 0.6 0.4 60 30 150 0.6 0.4 30 0.2 180 0 0.2 180 0 210 330 240 300 270 210 330 Precoder 1 240 Precoder 1 270 300 Layer 2 RBS Figure 3-9. Example of spatial multiplexing (SU-MIMO) LZT 123 9417 R1A 2009 Ericsson - 41 -
The same principle is illustrated in Figure 3-10 and Figure 3-11, where we can see how two different precoders give rise to different antenna patterns that exploit the multi path propagation geometry in two different ways as the UE moves. This is an example of SDM (Spatial Division Multiplexing) and is in LTE referred to as Single- User MIMO (SU-MIMO) 1 W = 1 1 1 Layer 2 Layer 2 Layer 1 RBS Figure 3-10. Example of spatial multiplexing DL SU-MIMO (SDM). W 1 = j 1 j Layer 2 Layer 2 Layer 1 RBS Figure 3-11. Example of spatial multiplexing DL SU-MIMO (SDM). - 42 - Ericsson 2009 LZT 123 9417 R1A
3 Precoding and Spatial Multiplexing Figure 3-12 shows how spatial multiplexing can be used in order to separate different users. This is an example of SDMA (Spatial Division Multiple Access) and is in LTE referred to as Multi-User MIMO (MU-MIMO). 1 W = j 1 j Layer 2 Layer 1 Layer 2 RBS Figure 3-12. Example of spatial multiplexing - UL MU-MIMO (SDMA). The spatial domain includes however more than the angular information that is illustrated in Figure 3-7 to Figure 3-12. It includes polarization and phase information. All these properties contribute to the spatial characteristics of the radio channel. This can be illustrated with vectors. The spatial multiplexing is performed in the vector space, where the vectors include polarization and carrier phase information. A UE reception may for instance be in a fading peak for one layer and one weight vector in the precoder matrix but in a fading dip for another layer with the same weight vector. The second layer may be in a fading peak for the second weight vector in the precoder matrix. So, when MU-MIMO is used, the system will try to find multiple UEs (UE antenna ports) that are orthogonal (or at least sufficiently well separated) in the vector space in either one or a combination of the following dimensions: azimuth angle (horizontal antenna diagram) polarization fading (carrier wave phase) LZT 123 9417 R1A 2009 Ericsson - 43 -
The UE signals will be separated in the receiver by applying the precoder on the receiver antenna ports. Performed in the vector space The vector space includes: Polarization Antenna radiation pattern Fading geometry Figure 3-13. Spatial multiplexing. When SU-MIMO is used, the system will try to find a precoder matrix that gives orthogonal (or at least sufficiently well separated) streams in the vector space after the receiver has performed its antenna demapping (applying the receiver precoder on its receiver antenna ports). - 44 - Ericsson 2009 LZT 123 9417 R1A
4 MIMO in WCDMA 4 MIMO in WCDMA OBJECTIVES: On completion of this chapter the student will be able to: Describe MIMO in WCDMA Explain Tx diversity in WCDMA Explain spatial multiplexing in WCDMA Describe the UE feedback (PCI) Describe the configuration of MIMO in WCDMA Figure 4-1. Objectives. LZT 123 9417 R1A 2009 Ericsson - 45 -
Intentionally Blank - 46 - Ericsson 2009 LZT 123 9417 R1A
4 MIMO in WCDMA MIMO IN WCDMA WCDMA RAN P7FP supports precoder-based 2x2 SU-MIMO as an optional feature. This feature allows for transmission of one or two data streams of HS-DSCH from 2 antenna arrays and reception on 2 antenna arrays in the downlink. Existing dual polarized antennas can be reused. 2x2 MIMO provides bit rates up to 28 Mbps (twice as much as without MIMO) and it also delivers an improved spectral efficiency and capacity. 64QAM Downlink 3.6 Mbps 14 Mbps 21 Mbps 28 Mbps 42 Mbps 15 codes 2x2 MIMO Both Multi Carrier 4x4 MIMO Higher Modulation Combinations Uplink 20-40 Mbps 12 Mbps 5.8 Mbps 1.4 Mbps Multi Carrier 16QAM 2 ms TTI 80-160 Mbps 0.384 Mbps Higher Speed, Lower cost per GByte Figure 4-2. HSPA speed evolution. Other channels, such as common channels, non-mimo channels and Primary and Secondary SCH may use Tx Diversity. P&S-SCH uses TSTD (Time Switched Transmit Diversity), while the common channels, non-mimo channels use STTD (Space Time Transmit Diversity). LZT 123 9417 R1A 2009 Ericsson - 47 -
Precoder based 2x2 MIMO tx for HS-DSCH P-SCH and S-SCH use TSTD (Time Switched Transmit Diversity) Other channels use STTD (Space Time Transmit Diversity) Switching between dual and single stream MIMO per TTI Precoding Control Info (PCI) is fed back from UE RBS may override PCI Single stream MIMO similar to Tx diversity Single stream MIMO precoder updated per TTI Tx diversity weight updated per slot Figure 4-3. MIMO in WCDMA. However, the CL Tx Diversity cannot be used together with the fractional DPCH, so it is not used in Ericsson products. Switching between MIMO and Tx Diversity for the HS-DSCH requires RRC signaling, while switching between dual and single stream MIMO can be done per TTI without RRC signaling. MIMO 28 Mbps TX RX MIMO for new terminals MIMO Improves speed & capacity TX TxDiv RX TxDiv for legacy terminals Improves speed at cell edge 14 Mbps Today TX RX Higher speed, Longer range or Deeper indoor coverage Figure 4-4. MIMO and tx diversity. With MIMO, two parallel data streams can be transmitted to one UE on the same set of HS-PDSCH codes, instead of one stream only as is the case with Rel-6 HS-PDSCH. This means that the peak bit rate for HS-DSCH can be doubled. Path loss / Distance from RBS - 48 - Ericsson 2009 LZT 123 9417 R1A
4 MIMO in WCDMA Transmission of two parallel data streams (two parallel transport blocks) is denoted dual-stream transmission, while transmission of one stream only is denoted single-stream transmission. Singlestream transmission is quite similar to TX diversity with closed loop mode 1, but with the difference that the antenna weighting is updated once per TTI instead of once per slot for the TX diversity case. The weighting of the streams and distribution over the two antennas is called precoding, and ensures that the HS-PDSCH power on the two antennas is the same, independent on if singlestream or dual-stream transmission is made. The choice of precoder is based on feedback from the UE. This feedback includes CQI (Channel Quality Indicator) and PCI (Precoding Control Info). The PCI indicates the best precoding matrix (out of four possible matrices) as estimated by the UE from pilot measurements. The P-CPICH is transmitted on antenna branch 1. An altered version of P-CPICH is transmitted on antenna branch 2. See Figure 4-5. Scheduler HARQ Turbo encoder Precoder W Modulation w 11 Spreading Scrambling w 12 P-CPICH 1 DPCH, CCH... & TFRC selection HARQ Turbo encoder Modulation Spreading Scrambling w 22 w 21 P-CPICH 2 or S-CPICH DPCH, CCH... w w w w 11 12 11 12 * w21 w 22 w w 21 22 = 0 w 11 w 12 w 21 w 22 1 1 j 2 e CQI φ 1 e Calc precoding matrix jφ, φ (the two streams are orthogonal) Precoding control info (PCI) from UE Single- / dual-stream CQI feedback from UE { π / 4, 3π / 4, 5π / 4, 7π / 4} The PCI indicates the UE s preferred value of w 12, other three weights then known Figure 4-5. Precoding in WCDMA. The P-CPICH transmission on the two antenna branches enables the UE to estimate the spatial properties of the radio channel. The UE will select the pre-coder matrix that is estimated to give the best possible throughput and report it to the base-station in a PCI (Pre-coder Control Indicator) on HS-DPCCH. The RBS may override the recommended PCI. LZT 123 9417 R1A 2009 Ericsson - 49 -
If the radio channel happens not to support dual stream transmissions, the UE will report a PCI indicating that Single Stream MIMO should be used. Single Stream MIMO is similar to Closed Loop Tx Diversity, but with the difference that in CL Tx Div, the chosen PCI is not echoed back to the UE and that the Tx Diversity is updated per slot (667μs) while Single Stream MIMO is updated per TTI (2ms). In order to use MIMO, both TX diversity and support of Enhanced layer 2 are required. Also, the UEs need to be MIMO capable and the RBS must have one power amplifier per antenna branch. MIMO is an optional feature both for RAN and UE in 3GPP Rel-7. Only UEs belonging to HS-DSCH UE categories 15, 16, 17, 18, 19 and 20 support this feature. This feature gives higher DL speeds of up to 28 Mbps for RABs using Enhanced uplink in uplink and HSDPA in downlink. HS-DSCH category Maximum codes L1 peak rates Modulation MIMO Category 1 5 1.2 16QAM Category 2 5 1.2 16QAM Category 3 5 1.8 16QAM Category 4 5 1.8 16QAM Category 5 5 3.6 16QAM Category 6 5 3.6 16 QAM Category 7 10 7.3 16QAM Category 8 10 7.3 16QAM Category 9 15 10.2 16QAM Category 10 15 14.0 16QAM Category 11 5 0.9 QPSK Category 12 5 1.8 QPSK Category 13 15 17.6 64QAM Category 14 15 21.1 64QAM Category 15 15 23.4 MIMO Category 16 15 28 MIMO Category 17 15 23.4 or 17.6 MIMO or 64QAM Category 18 15 28 or 21.1 MIMO or 64QAM Category 19 15 35.3 MIMO and 64QAM Category 20 15 42.2 MIMO and 64QAM Figure 4-6. HSPA terminal categories. In P7FP, 64-QAM cannot be used simultaneous with MIMO. To reach the highest bit rates SPB3 HW is required in the RNC. Also, core network support is required for the higher bitrates. - 50 - Ericsson 2009 LZT 123 9417 R1A