LTE Channel State Information (CSI)

LTE Channel State Information (CSI) Presented by: Sandy Fraser, Agilent Technologies

Agenda Channel State Information (CSI) different forms and definitions Channel Quality Information, Pre-Coding Matrix Indicator, Rank Indicator Terminology CSI controls, formats and reports Test Summary Examine one test each for wideband CQI, frequency selective (subband) CQI, PMI and RI What Agilent can do for CSI test!! Summary This presentation relies on the listener understanding a little MIMO, so we will revise a little as we go through the talk 2

The whole CSI concept Channel conditions change FACT!! 1) A UE moves from one area of good reception to an area with bad reception 2) A UE is moving along a street with deep/wide fades and highly variable radio conditions 3) We are sat at a café downloading our mail when a delivery truck parks in our line of sight to the enb In all these case CSI takes care of the UE s movement to a more robust coding scheme, less puncturing, lower modulation depth, different allocation in either frequency or time. As in life, success is all in timing! 3

Terminology I Spatial Multiplexing Transmit Diversity Channel (Channel) Rank Correlation Condition Number The process of transmitting data from multiple antennas on the same frequency at the same time Transmission of common data, but modified in some way, on more than one antenna The entire route, from transmission to reception, including all the analog & RF circuits & antennas, that could introduce unwanted coupling or distortion The number of useable data stream (layers) in a multi-antenna radio system A measure of the similarity between different signals (after the receiver antennas) A short term measure of the increase in SNR needed to recover a spatially multiplexed signal 4

Terminology II Codeword (Transmission) Layer Precoding Codebook Closed Loop MIMO Beamforming Beamsteering The input data after basic adaptation from the payload With spatial multiplexing, it is synonymous with a stream The process of cross coupling the signals before transmission (used in closed loop operation) to equalize the demodulated performance of the layers The look-up table of cross coupling factors used for precoding; shared by the mobile and base-station A mechanism used to continuously adapt the transmitted signal to suit the channel characteristics, using the precoder The process of cross coupling the signals at transmitter (or receiver) to adapt to the channel. LTE precoding is one example of doing this When beamforming with phased array, it is the process of tracking the movement of the mobile 5

Agenda Channel State Information different forms and definitions Channel Quality Information, Pre-Coding Matrix Indicator, Rank Indicator Terminology CSI controls, formats and reports Test Summary Examine one test each for wideband CQI, frequency selective (subband) CQI, PMI and RI What Agilent can do for CSI test!! Summary 6

Information required by UE to transmit/receive UE s need to know a lot of information before sending or receiving data Uplink When the UE can transmit and on which resources Which modulation, transport block size and redundancy version to use Adjustments to align timing with enb Downlink When the UE should listen for DL data. DL data may not be contiguous in frequency Which modulation, transport block size and redundancy version were used to transmit this data Is this downlink spatially multiplexed Whether to hop the PUSCH or not Power level For Spatially multiplexed DL what pre-coding has been applied Which HARQ process does this data belong to ALL Transmit of this new information block or is re-transmit send from the enb to Isthe this UE new on the data Downlink or re-transmitted Control Information data (DCI) NACK d blocks 7

Downlink Control Information (DCI) formats DCI Format Payload Usage 0 UL Assignments RB Assignments, TPC, MCS, PUSCH hopping flag 1 DL Assignments RB Assignments, TPC, HARQ, MCS 1A DL Assignments (compact) RB Assignments, TPC, HARQ, MCS, RA 1B DL Assignments (compact with pre-coding) RB Assignments, TPC, HARQ, MCS, TPMI, PMI 1C DL Assignments (VERY compact) RB Assignments 1D DL Assignments (compact with pre-coding and power offset) Multi user MIMO RB Assignments, TPC, HARQ, MCS, TPMI, DL Power offset 2 DL Assignments for closed loop MIMO RB Assignments, TPC, HARQ, MCS, pre-coding 2A DL Assignments for open loop MIMO RB Assignments, TPC, HARQ, MCS, pre-coding 2B DL Assignments for dual layer TM8 beamforming RB Assignments, TPC, HARQ, MCS, pre-coding 2C DL Assignments for dual layer TM9 8 layer non codebook multiplexing (Rel10) 3 TPC commands for PUSCH and PUCCH with 2 bit power adjustments 3A TPC commands for PUSCH and PUCCH with single bit power adjustments 4 UL Assignments for up to 4 layers, 2 per codeword and pre-coding (Rel10) RB Assignments, TPC, HARQ, MCS, pre-coding Power control, e.g. USER1, USER2, USER.etc using TPC-PUCCH-RNTI and TPC-PUSCH-RNTI Power control, e.g. USER1, USER2, USER.etc using TPC-PUCCH-RNTI and TPC-PUSCH-RNTI RB Assignments, TPC, HARQ, MCS, pre-coding 8

DCI example N6061A Protocol logging 9

HARQ Link Adaptation Retransmissions of a particular HARQ process use the same modulation and coding scheme as the initial transmission. Each subsequent retransmission simply reduces the effective code rate through incremental redundancy there are 4 redundancy versions for LTE Link adaptation (AMC: adaptive modulation and coding) with various modulation schemes and channel coding rates can be applied to the shared data channel. AMC optimises the transmission performance of each UE while maximizing the system throughput. If we use too low a modulation depth e.g. QPSK during good radio conditions, then we are utilizing more bandwidth (for a given desired data rate) than we need to If we use too high a modulation depth in poor conditions, we end up with too many re-transmissions Either way we are not making efficient use of the resources available Channel STATE Indicator, which includes Channel Quality Indicator (CQI) is the means by which the channel conditions are reported to the enb to optimise AMC process. 10

LTE 3GPP Channel Quality Indictor (CQI) 36.213 section 7.2 CQI reports can be Wideband or per sub-band Semi static, Higher Layer Configured or UE selected single or multiple sub-bands CQI only, or CQI plus Pre-coding Matrix Indicator (PMI) / Rank Indicator (RI) Transmitted on PUCCH for sub-frames with no PUSCH allocation or PUSCH with or without scheduling grant or if no UL-SCH Depends on spatial multiplexing Reports can be periodic or aperiodic (when signaled by DCI format 0 with CQI request field set to 1) The enb need not necessarily use the CQI reported from the UE CQI index modulation coding rate x 1024 0 out of range efficiency 1 QPSK 78 0.1523 2 QPSK 120 0.2344 3 QPSK 193 0.3770 4 QPSK 308 0.6016 5 QPSK 449 0.8770 6 QPSK 602 1.1758 7 16QAM 378 1.4766 8 16QAM 490 1.9141 9 16QAM 616 2.4063 10 64QAM 466 2.7305 11 64QAM 567 3.3223 12 64QAM 666 3.9023 13 64QAM 772 4.5234 14 64QAM 873 5.1152 15 64QAM 948 5.5547 36.213 Table 7.2.3-1: 4-bit CQI Table 11

Channel State Indication CSI on Uplink Channel Information (UCI) 36.213 Section 7.2 Transmission Mode Payload 1. Single-antenna port; port 0 UE selected sub-band CQI + wide-band CQI or Higher Layer Configured wide-band and sub-band CQI, no PMI 2. Transmit diversity UE selected sub-band CQI + wide-band CQI or Higher Layer Configured wide-band and sub-band CQI, no PMI 3. Open-loop spatial multiplexing 4. Closed-loop spatial multiplexing UE selected sub-band CQI + wide-band CQI or Higher Layer Configured wide-band and sub-band CQI, no PMI Wide-band CQI per codeword + PMI for each sub-band or UE selected sub-band and wide-band CQI per codeword + PMI or Higher Layer Configured wide-band and sub-band CQI + PMI 5. Multi-user MIMO Higher Layer Configured wide-band and sub-band CQI + PMI 6. Closed-loop Rank=1 pre-coding Wide-band CQI per codeword + PMI for each sub-band or UE selected sub-band and wide-band CQI per codeword + PMI or Higher Layer Configured wide-band and sub-band CQI + PMI TM7, 8, 9 not listed 12

UCI on the PUCCH or PUSCH Physical Uplink Control Channel (PUCCH) or Physical Uplink Shared Channel carries the Uplink Control Information CQI and ACK/NACK, and also scheduling requests Format Bits per sub-frame Payload Mod n 1 N/A No Ack/Nack, only SR N/A 1a 1 SISO Ack/Nack BPSK 1b 2 MIMO Ack/Nack QPSK 2 20 CSI, no Ack/Nack QPSK 2a * 21 CSI + SISO Ack/Nack B/QPSK 2b * 22 CSI + MIMO Ack/Nack B/QPSK The number and position of Demodulation Reference Signal symbols will vary depending on format * For normal CP only 13

UCI example N6061A Protocol logging Periodic CQI report combined with ACK/NACK reporting Aperiodic CQI report with PMI and RI 14

The mandated CSI tests 36.521 section 9, 36.101 section 9 There are 18 CSI tests 10 for CQI testing (both wide-band and sub-band CQI) 6 for PMI testing 2 for RI testing Almost all are COMPARATIVE tests several stages with varying conditions results compared to ensure throughput gain Almost all require AWGN and Fading NONE are truly representative of real world conditions. All have separate FDD and TDD sections 7 are currently defined for Release 10 in 36.101 requirements (CSI reference symbols), but are not defined in the test procedures 36.521 These standards change regularly this section last updated May 2012 16

36.521 Section 9 CSI Conformance Tests Most requirements are tested using faded DL channels Most are comparative tests, accomplished in several stages Most employ fixed or minimally varied transmission conditions Test Title and 3GPP 36.521 test reference Channel SNR options, test count Mode Description CQI reporting under AWGN 9.2.1 CQI reporting under AWGN 9.2.2 AWGN (1 x 2) 2,2 PUCCH 1-0 AWGN (2 x 2) 2,2 PUCCH 1-1 Comparison of BLER using CQImedian +/-1 values Comparison of BLER for each codeword using CQImedian +/-1 values CQI reporting under AWGN 9.2.3 Defined in 36.101 (requirements), not in 36.521 (procedures) Defined in 36.101 (requirements), not in 36.521 (procedures) PUCCH 1-1 CSI Reference Symbols placeholder R10 feature 17

Summary of test cases continued - CQI 18 Test Title and 3GPP 36.521 test reference CQI Frequency-selective scheduling 9.3.1.1 CQI Frequency-selective scheduling 9.3.1.2 CQI Frequency nonselective scheduling 9.3.2.1 CQI Frequency nonselective scheduling 9.3.2.2 CQI Frequency-selective interference 9.3.3 Sub band size 6RB CQI UE-selected sub-band 9.3.4.1 Sub-band size 3RB CQI UE-selected sub-band 9.3.4.2 Sub-band size 6RB Channel 3GPP 36.101 Clause B.2.4 with specific fading conditions Defined in 36.101 (requirements), not in 36.521 (procedures) SNR options, test count Mode 2,2 PUSCH 3-0 Defined in 36.101 (requirements), not in 36.521 (procedures) EPA5, High 2,2 Defined in 36.101 (requirements), not in 36.521 (procedures) 3GPP 36.101 Clause B.2.4 with specific fading conditions 3GPP 36.101 Clause B.2.4 with specific fading conditions 3GPP 36.101 Clause B.2.4 with specific fading conditions Defined in 36.101 (requirements), not in 36.521 (procedures) PUSCH 3-1 PUCCH 1-0 on PUSCH to avoid CQI and ACK collisions PUCCH 1-1 2,1 PUSCH 3-0 2,2 PUSCH 2-0 2,2 PUSCH 2-0 to avoid CQI and ACK collisions Description Throughput with enb random sub-band allocation, then test with UE reported sub-band allocation. Differential minimum throughput gain ratio. CSI Reference Symbols placeholder R10 feature Compares Throughput using UE reported CQI against fixed CQImedian CSI Reference Symbols placeholder R10 feature Throughput with enb random sub-band allocation, then test with UE reported sub-band allocation. Differential minimum throughput gain ratio. Throughput with enb random sub-band allocation, then test with UE reported sub-band allocation. Differential minimum throughput gain ratio. Throughput with enb random sub-band allocation, then test with UE reported sub-band allocation. Differential minimum throughput gain ratio.

Summary of test cases continued - PMI Test Title and 3GPP 36.521 test reference Channel SNR options, test count Mode Description Single PMI reporting 9.4.1.1 Single PMI reporting 9.4.1.2 Single PMI reporting 9.4.1.3 Multiple PMI reporting 9.4.2.1 Multiple PMI reporting 9.4.2.2 Multiple PMI reporting 9.4.2.3 EVA5, Low 2x2 1,1 PUSCH 3-1 EVA5, Low 4x2 1,1 PUCCH 2-1 Defined in 36.101 (requirements), not in 36.521 (procedures) Defined in 36.101 (requirements), not in 36.521 (procedures) PUSCH 3-1 EPA5, Low 2x2 1,1 PUSCH 1-2 EVA5, Low 4x2 1,1 PUSCH 2-2 Defined in 36.101 (requirements), not in 36.521 (procedures) Defined in 36.101 (requirements), not in 36.521 (procedures) PUSCH 3-1 Compares random pre-coding matrix reported TP against UE reported precoding matrix TP. Compares random pre-coding matrix reported TP against UE reported precoding matrix TP. CSI Reference Symbols placeholder R10 feature Compares random pre-coding matrix reported TP against UE reported precoding matrix TP. Compares random pre-coding matrix reported TP against UE reported precoding matrix TP. CSI Reference Symbols placeholder R10 feature 19

Summary of test cases continued - RI Test Title and 3GPP 36.521 test reference Channel SNR options, test count Mode Description Rank Indicator reporting 9.5.1.1 Rank Indicator reporting 9.5.1.2 Rank Indicator reporting 9.5.2.1 Rank Indicator reporting 9.5.2.2 EPA5, Low and high 2x2 1, 3 EPA5, Low and high 2x2 1, 3 Defined in 36.101 (requirements), not in 36.521 (procedures) Defined in 36.101 (requirements), not in 36.521 (procedures) Defined in 36.101 (requirements), not in 36.521 (procedures) Defined in 36.101 (requirements), not in 36.521 (procedures) PUCCH 1-1 FDD Only PUCCH 3-1 TDD Only FDD only TDD only Compares TP with fixed rank, vs reported rank for 3 separate channel and rank conditions. Compares TP with fixed rank, vs reported rank for 3 separate channel and rank conditions. CSI Reference Symbols placeholder R10 feature CSI Reference Symbols placeholder R10 feature 20

Wideband CQI Test With AWGN (PUCCH format 1.0), 9.2.1 Parameter Unit Test 1 Test 2 Bandwidth MHz 10 PDSCH transmission mode 1 Downlink power A db 0 allocation B db 0 Propagation condition and antenna configuration AWGN (1 x 2) SNR (Note 2) db 0 1 6 7 ( ) Iˆ j or db[mw/15khz] -98-97 -92-91 db[mw/15khz] -98-98 Max number of HARQ transmissions 1 Physical channel for CQI reporting PUCCH Format 2 PUCCH Report Type 4 Reporting periodicity ms NP = 5 cqi-pmi-configurationindex 6 Note 1: Note 2: ( j) N oc Reference measurement channel according to Table A.4-1 with one sided dynamic OCNG Pattern OP.1 FDD as described in Annex A.5.1.1. For each test, the minimum requirements shall be fulfilled for at least one of the two SNR(s) and the respective wanted signal input level. PUCCH 1-0 static test (36.101 [10] Table 9.2.1-1) 21

Wideband CQI Test With AWGN (PUCCH format 1.0), 9.2.1 Setup with conditions stated and measure the median value of CQI 90% of all 2000 CQI results obtained must be within +/- 1 of this median value Take this median CQI-1 value and measure BLER which must be less than 10%. Take the median CQI +1 value and measure the BLER which must be greater than 10%. If the UE fails this test using the first SNR value (0 db), then the test sequence can be repeated using the second value (1 db). The UE must pass at least one of these two tests. The test is then repeated for the SNR of 6dB, and if necessary 7dB. Med CQI-1 Test part 2 with lower than optimal data flow results in low BLER (less than 10%) Med CQI Med CQI+1 Establish Median CQI in test part 1 Test part 3 with higher than optimal data flow results in high BLER (greater than 10%) 22

Frequency Selective CSI What if? What if I have been allocated the resources in RED BUT my UE measures the PURPLE area to be more suitable for the measured channel conditions? What happens if the conditions have changed by the time the UE is moved to these RB s? This is the purpose of sub-band (frequency selective) CSI testing Time Frequency 23

Frequency selective (sub-band) scheduling CQI test with fading, 9.3.1.1 Parameter Unit Test 1 Test 2 Bandwidth MHz 10 MHz Transmission mode 1 (port 0) SNR (Note 3) db 9 10 14 15 ( ) Iˆ j or db[mw/15khz] -89-88 -84-83 ( j) N oc db[mw/15khz] -98-98 Propagation channel 3GPP 36.101 Clause B.2.4 with specific fading conditions Correlation Full Reporting interval ms 5 CQI delay ms 8 Reporting mode PUSCH 3-0 Max number of HARQ transmissions 1 Note 1: Note 2: Note 3: If the UE reports in an available uplink reporting instance at subframe SF#n based on CQI estimation at a downlink subframe not later than SF#(n-4), this reported subband or wideband CQI cannot be applied at the enb downlink before SF#(n+4) Reference measurement channel according to Table A.4-4 with one/two sided dynamic OCNG Pattern OP.1/2 FDD as described in Annex A.5.1.1/2 For each test, the minimum requirements shall be fulfilled for at least one of the two SNR(s) and the respective wanted signal input level. Sub-band test for single antenna transmission (FDD) (36.101 [10] Table 9.3.1.1.1.3-1) 24

Frequency selective (sub-band) scheduling CQI test with fading, 9.3.1 Gather 2000 CQI reports The sub-band differential CQI offset level of 0 shall be reported at least a % of the time but less than b % for each sub-band Parameter Test 1 Test 2 a [%] 2 2 b [%] 55 55 γ 1.1 1.1 One sub-band may be different size to others this one is not used because it could skew the throughput results Ignoring reports from the UE, the 2 nd stage of the test allocates random subbands to the UE and tests throughput. The 3 rd stage of the test uses the highest ranking sub-bands reported by the UE The ratio of stage 2 and 3 should represent a throughput gain of more than 10% AND the BLER must be greater than 5%. If the UE fails this test using the first SNR value (9 db), then the test sequence can be repeated using the second value (10 db), the UE must pass at least one of these two tests. The test is then repeated using the SNR values 14 (and if necessary) 15dB. 25

So Tell (or Remind) Me How does MIMO work? 1: Consider a moment in time, at a single frequency, and model the channel as a box with fixed components inside: A B C D If we add two completely different signals at A and B, they ll get mixed together, but in a precisely defined way, dependant on the values of Z1- Z4 MIMO is used uncouple signals on twisted pairs 2: Send a training signal first, that s unique to A and to B. Measure what comes out at C and D and therefore how they got coupled. [If you know how they get coupled, you can work out how to uncouple them] 3: Everything going into the box will be coupled the same way, so you apply what you found to the real data you want to sent 26

and when does it not work? Noise & interference always limit the modulation we use. With MIMO, there is an ADDITIONAL factor how well can you uncouple the signals measured by the Condition Number of the channel matrix A B C D Extreme example: If all the Z s are the same, both outputs are the same. This is a keyhole channel, which does not support spatial multiplexing (rank =1) For every db increase in condition number, you may need a db increase in the SNR 27

Why Precode (cross couple) the SM signal? No precoding the layer performance is unbalanced Precoded to achieve similar performance for both layers 28

Precoding Matrix Index definition 3GPP TS 36.211 Table 6.3.4.2.3-1 Deals with FDD case Only 3 (2 for TM4) choices for spatial multiplexing (16 for the 4 layer case) For single data stream transmission, the precoding produces beamsteering (with 4 antennas) Subband PMI reporting can be configured down to the resource block level 29

PMI Testing, 9.4.1.1 Note 1: Note 2: Parameter Unit Test 1 Bandwidth MHz 10 Transmission mode 6 Propagation channel EVA5 Precoding granularity PRB 50 Correlation and antenna configuration Low 2 x 2 A db -3 Downlink power allocation PMI test for single layer (FDD) (36.101 [10] Table 9.4.1.1.1.3-1) ( j) N oc B db -3 db[mw/15khz] -98 Reporting mode PUSCH 3-1 Reporting interval ms 1 PMI delay (Note 2) ms 8 Measurement channel R.10 FDD OCNG Pattern OP.1 FDD Max number of HARQ transmissions 4 Redundancy version coding sequence {0,1,2,3} For random precoder selection, the precoder shall be updated in each TTI (1 ms granularity) If the UE reports in an available uplink reporting instance at subrame SF#n based on PMI estimation at a downlink SF not later than SF#(n-4), this reported PMI cannot be applied at the enb downlink before SF#(n+4). 30

PMI Testing, 9.4.1 The first stage of the test is performed in order to establish the value SNR(rnd). This is the Signal to Noise Ratio used during the second and third stages of the test. 36.101 Annex G.5.2 specifies how to establish the value SNR(rnd), by adjusting the SNR until the throughput is settled between 58% and 62% of the calculated maximum throughput t(rnd). The second stage of the test is performed using random pre-coding The third stage repeats stage 2 but using UE reported PMI values. Throughput results are obtained using these two different conditions, and the throughput ratio (γ ) is expressed as precoding gain A pass is achieved if the ratio γ is exceeded. Table 6.6-13. Minimum requirement (FDD) (36.101 [10] Table 9.4.1.1.1.3-2) Parameter Test 1 γ 1.1. 31

Rank Index Only certain channel models are suitable for MIMO If MIMO is used when the channel can only support 1 stream of data the resulting throughput will be poor and resources wasted If MIMO is NOT used when the channel CAN support more than one stream, then the throughput will be low and resources wasted. 32

Rank Indication (RI) Testing, 9.5.1.1 Parameter Unit Test 1 Test 2 Test 3 Bandwidth MHz 10 PDSCH transmission mode 4 Downlink power A db -3 allocation B db -3 000011 for fixed RI = 1 CodeBookSubsetRestriction 010000 for fixed RI = 2 bitmap 010011 for UE reported RI Propagation condition and antenna configuration 2 x 2 EPA5 Antenna correlation Low Low High RI configuration RI test (FDD) (36.101 [10] Table 9.5.1.1.3-1) Fixed RI=2 and follow RI Fixed RI=1 and follow RI Fixed RI=2 and follow RI SNR db 0 20 20 ( j) N oc db[mw/15khz] -98-98 -98 ( ) Iˆ j or db[mw/15khz] -98-78 -78 Maximum number of HARQ transmissions 1 Reporting mode PUCCH 1-1 (Note 4) Physical channel for CQI/PMI reporting PUCCH Format 2 PUCCH Report Type for CQI/PMI 2 Physical channel for RI reporting PUSCH (Note 3) PUCCH Report Type for RI 3 Reporting periodicity ms NP = 5 PMI and CQI delay ms 8 cqi-pmi-configurationindex 6 ri-configurationind 1 Note 1: Note 2: Note 3: Note 4: If the UE reports in an available uplink reporting instance at subframe SF#n based on PMI and CQI estimation at a downlink subframe not later than SF#(n- 4), this reported PMI and wideband CQI cannot be applied at the enb downlink before SF#(n+4). Reference measurement channel according to Table A.4-1 with one sided dynamic OCNG Pattern OP.1 FDD as described in Annex A.5.1.1. To avoid collisions between RI reports and HARQ-ACK it is necessary to report both on PUSCH instead of PUCCH. PDCCH DCI format 0 shall be transmitted in downlink SF#4 and #9 to allow periodic RI to multiplex with the HARQ-ACK on PUSCH in uplink subframe SF#8 and #3. The bit field for precoding information in DCI format 2 shall be mapped as: -For reported RI = 1 and PMI = 0 >> precoding information bit field index = 1 -For reported RI = 1 and PMI = 1 >> precoding information bit field index = 2 -For reported RI = 2 and PMI = 0 >> precoding information bit field index = 0 33

Rank Indication (RI) Testing, 9.5.1 Test stage (a) establishes the value t(fix), Using the CodeBookSubsetRestriction for fixed Rank (1 or 2), the system simulator responds with UL grants to the UE based on the CQI, RI, and PMI reports from the UE. For test stage (b), the UE is then told to use the CodeBookSubsetRestriction as for UE reported RI shown in table 6.6-14, along with all the other parameters to establish t(reported). The ratio of the two throughput values γ obtained from the two test stages should satisfy the requirements shown in table 6.6-15 Table 6.6-15. RI minimum requirements (FDD) (36.101 [10] Table 9.5.1.1.3-2) Parameter Test 1 Test 2 Test 3 γ 1 N/A 1.05 N/A γ 2 1.0 N/A 1.1 The first test should give very similar throughput values for the two test stages. Due to the low SNR value there will be little or no improvement expected. The second test should show a modest throughput improvement, but will still be restricted due to the use of R1 for both test stages, while test three will show the highest improvement because of the highest SNR and use of fixed R2 for the first stage of the test. 34

AWGN and OCNG Required for most section 7,8,9 tests AWGN - Settable SNR for each RF1, RF2, MIMO or Normal settings for channel mode Indicated values for NoC and Noise Amplitude Amp > AWGN OCNG defined in 36.521-1 section A.5 Fills any un-used RB s with OCNG Mode > BSE > Func > OCNG 36

Closed Loop TM4 and TM6 testing BSE>mode setup>more>rrc>tm Mode BLER/Tput Testing 64QAM MCS 17-25 AUTO QPSK MCS 0-9 AUTO 1 3 AUTO 0 Supports Test Mode and E2E Testing DL MCS 16QAM MCS 10-16 2 RI 2 PMI 1 Open Loop and Closed Loop Testing Display CQI/RI/PMI reported information CHANNEL EMULATOR CQI PMI RI 37

Comprehensive Throughput Reporting DL and UL throughput graphs and values Average throughput, BLER, ACK, NACK and StatDTX counts New tab for channel state information (CQI, PMI, RI etc) 38

CSI Reporting CSI = channel state information includes CQI, PMI, RI New tab for channel state information (CQI, PMI, RI etc) Wideband and subband reports PMI, RI reports Periodic, Aperiodic reporting (depends on scenario and front panel settings 36.521 section 9 automatic reports and measurements 39

Differential CQI reported values 40

CQI reporting for 36.521 Section 9 CSI = channel state information includes CQI, PMI, RI Statistical CQI Performance The Statistical CQI Performance measurement is used as part of an RCT system to perform test cases in 36.521-1, section 9. Mode > BSE > Func > More > RCT > Statistical CQI Performance Median CQI This setting starts/initiates the collection of CQI reports from UE. Aperiodic CQI, Periodic CQI The scenario must contain the appropriate CQI Report Configuration (either periodic or aperiodic) in the RRC Setup message information. This enables the UE to generate the correct CQI reports. Mode > BSE > Func > More > CQI Median DL Allocation based on CQI Set the DL allocation to whatever the UE is reporting Choose from Wideband or Sub-band and random Sub-band Mode > BSE > Mode Setup > More > PHY > DL Resources > CQI Reports 41

CQI Control Loop Testing E6621A PXT (enodeb emulator) N9020A MXA Signal Analyzer Ext Trig 1 In N5182A MXG Vector Signal Generator Trig Out Pulse Trig In Sweep out Does the CQI feedback process act fast enough? 50ms noise bursts added to downlink 42

Testing CQI Using UE Reporting (Delayed ) drop in MCS in response to CQI report from UE 2 frame quantisation in response set by reporting interval Over-damped control loop response showing impact of reporting interval and CQI UE report averaging HARQ retransmissions occur throughout the noise burst 43

Testing CQI Using UE Reporting Random re-transmissions and RV s Periods of StatDTX (No ACK s or NACK s indicating no reception by UE or no report sent by UE 44

Summary LTE CSI testing is more involved than originally for W-CDMA 18 tests, FDD/TDD = 36, multi-steps = more than 100 test steps Testing requires very specific setting capability in test equipment Testing is largely static with fixed I_MCS, although using faded conditions this does not represent the real world use case. Real world type testing may be required to ensure end user satisfaction. Fixed channel testing will be required to debug persistent throughput or CSI reporting issues 45

Resources Webcast for more detailed description of analysis using VSA software http://www.eetimes.com/electrical-engineers/educationtraining/webinars/4211278/how-to-verify-the-data-in-your-lte-uplink-signal Agilent VSA site: www.agilent.com/find/vsa This webcast was recorded and will be available shortly along with the slides Everything related to the PXT network emulator including Radio (pre-) Conformance testing www.agilent.com/find/pxt Application notes, white papers, demonstrations, webcasts, training events, related products and MUCH more: www.agilent.com/find/mimo 46

Backup slides 47

BACKUP SLIDES - LTE - MIMO

Agenda Overview of Multi-antenna techniques LTE Terminology How MIMO works in LTE 49 LTE RF Design and Measurement Course

Multi-Antenna Techniques in LTE Just because there is more than one antenna, doesn t mean it s MIMO Diversity can usefully be combined with MIMO Spatial Multiplexing to improve performance A focus on the need to provide an increased DL data rate leads to an asymmetric system in LTE 50

System & Antenna Configurations Terms Input and Output Refer to the Channel SISO MISO Tx0 Rx Tx Rx SIMO Tx1 Tx Diversity, Beamforming MIMO Tx Rx0 Tx0 Rx0 Rx Diversity Rx1 Tx1 Rx1 Spatial Multiplexing 51

MIMO Spatial Multiplexing and Diversity Both Important, Different Objectives Multiple Antennas can be used in a variety of ways: Beamforming Transmit Diversity Receive Diversity Diversity techniques protect against fading, and improve coverage 53 LTE RF Design and Measurement Course

Double Diversity does not make MIMO Transmit Diversity + Receive Diversity = Spatial Multiplexing MISO plus MRC Tx0 Tx1 Data modified and repeated on second symbol (or subcarrier) Tx0 Tx1 MIMO Data only transmitted once Tx0 Tx1 Rx0 Rx1 54

MIMO Operation in LTE In the Downlink, it s normally like WLAN, the MIMO transmission is sent to a single mobile. Known as Single User MIMO In the Uplink, two mobiles are used together to create the MIMO signal. Known as Multi-User MIMO 55 LTE RF Design and Measurement Course

and when does it not work? Noise & interference always limit the modulation we use. With MIMO, there is an ADDITIONAL factor how well can you uncouple the signals measured by the Condition Number of the channel matrix A B Extreme example: If all the Z s are the same, both outputs are the same. This is a keyhole channel, which does not support spatial multiplexing (rank =1) For every db increase in condition number, you may need a db increase in the SNR 58

LTE Channel Training Signals The Reference Signals are what allow the receiver to calculate the channel coefficients. They NEVER overlap before they are transmitted R 0 R 0 One antenna port R 0 R 0 R 0 R 0 R 0 l 0 R 0 l 6 l 0 l 6 Resource element (k,l) R 0 R 0 R 1 R 1 Two antenna ports R 0 R 0 R 0 R 0 R 1 R 1 R 1 R 1 Not used for transmission on this antenna port Reference symbols on this antenna port R 0 R 0 R 1 R 1 l 0 l 6 l 0 l 6 l 0 l 6 l 0 l 6 R 0 R 0 R 1 R 1 R 2 R 3 Four antenna ports R 0 R 0 R 0 R 0 R 1 R 1 R 1 R 1 R 2 R 2 R 3 R 3 R 0 R 0 R 1 R 1 R 2 R 3 l 0 l 6 l 0 l 6 l 0 l 6 l 0 l 6 l 0 l 6 l 0 l 6 l 0 l 6 l 0 l 6 even-numbered slots odd-numbered slots even-numbered slots odd-numbered slots even-numbered slots odd-numbered slots even-numbered slots odd-numbered slots Antenna port 0 Antenna port 1 Antenna port 2 Antenna port 3 59

What makes a good channel for MIMO? h 00 Channel H T 0 R 0 1 0 A perfect MIMO channel: T 1 h 11 R 1 0 1 By simple observation it follows that R 0 = T 0 and R 1 = T 1 This is a case that creates double the capacity But suppose we create a simple static channel like this: How do we know if it will provide capacity gain? Channel H 0.8 0.2 0.3-0.9 60

The MIMO challenge Recovering the signal So is the earlier example good or bad for MIMO? R 0 = 0.8 T 0 + 0.3 T 1 R 1 = 0.2 T 0-0.9 T 1 Giving: Channel H 0.8 0.2 0.3-0.9 T 0 = 1.15 R 0 + 0.39 R 1 T 1 = 0.26 R 0-1.03 R 1 We can recover the original signal In fact any H matrix other than the unity matrix can be resolved PROVIDED there is no external or internal noise! Page 61 61

Why Precode (cross couple) the SM signal? No precoding the layer performance is unbalanced Precoded with 1,1,-1,1 similar performance for both layers 62

Precoding Matrix Index definition 3GPP TS 36.211 Table 6.3.4.2.3-1 Deals with FDD case Only 3 choices for spatial multiplexing (16 for the 4 layer case) For single data stream transmission, the precoding produces beamsteering (with 4 antennas) Subband PMI reporting can be configured down to the resource block level 63

Antenna influence on performance The dynamic condition number example did not isolate effects from different components, including the antenna In real life, the instantaneous channel matrix H is made up from the interaction of three components: The static 3D antenna pattern of the transmitter The dynamic multipath and Doppler characteristics of the radio channel The static 3D antenna pattern of the receiver The overall antenna contribution is the product of the transmit and receive antennas known as the channel correlation matrix 64

Real life performance Variation due to instantaneous correlation Variation in the frequency domain not shown Most macrocell activity takes place in this region Variation due to fading and variable interference 65

Summary MIMO Spatial Multiplexing is a powerful additional transmission scheme in the right conditions The list of 7 modes for DL transmission highlights how the ENB and UE will have to work together to choose which multi-antenna technique to use: LTE has seven different downlink transmission modes: 1.Single-antenna port; port 0 2.Transmit diversity 3.Open-loop spatial multiplexing 4.Closed-loop spatial multiplexing 5.Multi-user MIMO 6.Closed-loop Rank=1 precoding 7.Single-antenna port; port 5 SISO MISO MIMO no precoding MIMO with precoding MIMO - separate UE (for UL) MISO - beamsteering MISO beamsteering 66