Internal
06/22/16
LTE System Multiple Antenna Techniques
eRAN2.2 (MIMO and Beamf!min"#
HUAWEI TECHNOLOGIES CO., LTD.
Huawei Confdential
www.huawei.com
Training Tr aining Objectives
Afte! cmpletin" this cu!se$ yu %ill &e a&le t'
Understand the concepts relevant to the MIMO and Beamforming
Understand basic principle of MIMO and Beamforming
Refe!ences'
!"## T$ !6211% #h&sical 'hannels and Mod(lation
!"## T$ !621!% #h&sical la&er proced(res
!"## T$ !6!06% User )*(ipment +U), radio access capabilities
-#.% MIMO and Beamforming -eat(re .oc(mentation
Training Tr aining Objectives
Afte! cmpletin" this cu!se$ yu %ill &e a&le t'
Understand the concepts relevant to the MIMO and Beamforming
Understand basic principle of MIMO and Beamforming
Refe!ences'
!"## T$ !6211% #h&sical 'hannels and Mod(lation
!"## T$ !621!% #h&sical la&er proced(res
!"## T$ !6!06% User )*(ipment +U), radio access capabilities
-#.% MIMO and Beamforming -eat(re .oc(mentation
'ontents
Bac"!und and O)e!)ie% f the LTE MIMO Techniques
*!inciples and Applicatin f the MIMO Techniques
*!inciples and Applicatin f Beamf!min" Beamf!min"
Bac"!und f Multi+Antenna Techniques
-ift& &ears ago $hannon gave the maim(m efficienc& that a time and fre*(enc& comm(nication s&stem can achieve
S C = B × log 2 1 + ( bit / s ) N The rapid development of ireless comm(nications poses increasingl& higher re*(irement for s&stem capacit& and spectral efficienc& ario(s algorithms are invented s(ch as spreading the s&stem bandidth optimi3ing the mod(lation scheme or (sing comple code division m(ltiple access These methods are limited% Bandidth cannot be epanded indefinitel&4 mod(lation orders cannot increase indefinitel&4 channels beteen a '.M5 s&stem are not ideall& orthogonal 5nother dimension that is MIMO is invented to better (se the spatial reso(rce 5s epressed in the folloing e*(ation if m(ltiple antennas are (sed the capacit& is increased b& a m(ltiplication of the n(mber of antennas (sed
S C = B × log 2 1 + ( bit / s ) × M N
Ad)anta"es f Multi+Antenna Techniques
The T) s&stem improves s&stem performance for cell edge (sers and brings stable and reliable service eperience for (sers Therefore m(lti7antenna techni*(es can ma8e (se of the spatial reso(rce and increase the ireless transmission capacit& man& folds itho(t increasing the transmit poer and bandidth
5rra&
gain
Improved
s&stem coverage .iversit&
gain
Improved
s&stem capacit& $patial
m(ltipleing
gain Increased
rate 'o7channel interference
red(ction
pea8
Increased spectral efficienc&
'ontents
Bac"!und and O)e!)ie% f the LTE MIMO Techniques
*!inciples and Applicatin f the MIMO Techniques
*!inciples and Applicatin f Beamf!min"
*!inciples f the MIMO Techniques
MIMO is an important techni*(e in the T) s&stem. MIMO means (se of m(ltiple antennas at both the transmitter and receiver MIMO can better (tili3e the spatial reso(rce and increase spectral efficienc& achieving arra& gain diversit& gain m(ltipleing gain and interference rejection gain providing higher s&stem capacit& ider coverage and higher (ser rate
,lassificatin f MIMO Techniques
-ependin" n %hethe! the spatial channel inf!matin is used$ MIMO techniques a!e classified int pen+lp MIMO and clsed+lp MIMO. < Open7loop MIMO% The U) does not feed bac8 information the e=odeB is not informed of the U) sit(ation The protocols s(pport single7stream +TM2, or m(lti7stream +TM!, < 'losed7loop MIMO% The U) feeds bac8 information The gain has a positive correlation ith the acc(rac& of the feedbac8 information The protocols s(pport single7stream +TM>, or m(lti7 stream +TM6, 5t present the feedbac8 gran(larit& s(pported b& the reference signal in port 2 is large and closed7loop MIMO can hardl& achieve gains 'losed7loop MIMO re*(ires lo U) mobilit& 5t present the e=odeB cannot acc(ratel& estimate the U) movement speed ith an error of more than !0 8m/h -ependin" n the num&e! f simultaneusly t!ansmitted spatial data st!eams$ MIMO techniques a!e classified int spatial di)e!sity and spatial multiplein". These mdes a!e desc!i&ed in detail in the fll%in" pa"es.
MIMO Techni*(e
M(lti7antenna receive
MIMO Mode
?eceive diversit&
MU7MIMO
M(lti7antenna transmit
Open7loop transmit diversit& 'losed7loop transmit diversit& Open7loop spatial m(ltipleing 'losed7loop spatial m(ltipleing
-eat(re ist in -..
-eat(re ist in T..
U 275ntenna ?eceive .iversit&
U 275ntenna ?eceive .iversit&
U >75ntenna ?eceive .iversit&
U >75ntenna ?eceive .iversit&
U Interference ?ejection 'ombining
U Interference ?ejection 'ombining
U 22 MU7MIMO
U ;75ntenna ?eceive .iversit& U 22 MU7MIMO
U 2> MU7MIMO
U 2> MU7MIMO
22 MIMO >2 MIMO . >> MIMO
22 MIMO >2 MIMO
*!inciple f Multi+Antenna Recei)e MIMO
e?5=22 s(pports U 275ntenna ?eceive .iversit& and optional U >75ntenna ?eceive .iversit& and U ;75ntenna ?eceive .iversit& The folloing fig(re shos the bloc8 diagram of receive diversit& The U) (ses one antenna to transmit signals4 different U)s (se different time and fre*(enc& reso(rces The e=odeB (ses m(ltiple antennas to receive signals and combine the received signals to maimi3e $I=? therefore obtaining diversit& gain and arra& gain increasing the cell coverage and improving single7(ser capacit&
*!inciple f Multi+Antenna Recei)e MIMO
Mechanism f Si"nal ,m&inatin' 5n MM$) receiver (ses receive beamforming targeted at a U) The receiver adj(sts the combined eight and changes the direction of the major lobe and side lobe to maimi3e the $I=? of the received signals There are to combination algorithms for U receive diversit& Maim(m ratio combining +M?', and interference rejection combining +I?', can both obtain diversit& gain and arra& gain improving s&stem performance M?' and I?' are s(itable for environments ith different interference characteristics M?' receivers and I?' receivers are implementation of MM$) receivers in different scenarios
Specificatin f Multi+Antenna Recei)e MIMO
Adapti)e S%itch)e! Bet%een MR, and IR,
-or e=odeBs 22 I?' is optional If I?' is not selected an e=odeB (ses M?' If I?' is selected an e=odeB adaptivel& selects I?' or M?' depending on the c(rrent radio channel *(alit&
If there is separable strong colored interference the s&stem a(tomaticall& (ses I?' algorithm
If there is no separable strong colored interference the s&stem a(tomaticall& rolls bac8 to M?' algorithm
In U 22 MU7MIMO mode the e=odeB does not s(pport U Interference ?ejection 'ombining or U 275ntenna ?eceive .iversit&
In U >75ntenna ?eceive .iversit& mode the e=odeB s(pports U Interference ?ejection 'ombining
#rinciple of M(lti7User MIMO +MU7MIMO,
Theoreticall& the n(mber of virt(al MIMO (sers in the same ?B cannot eceed the n(mber of receive antennas of the e=odeB e=odeBs22 s(pport MU7MIMO 22 The folloing fig(re shos MU7MIMO 22
e=odeB22 ， The protocols s(pport a maim(m of MU7MIMO >>
M(lti75ntenna Transmit MIMO
The e=odeB s(pports m(lti7antenna transmission and the U) does not . 22 MIMO . >2 MIMO and . >A> MIMO are described ?@ defines nine m(lti7antenna transmission modes +TMs, The e=odeB adaptivel& selects one TM according to the channel condition and service re*(irement Supp!ted &y N. Name
U s e d b & - . . / T . .
1 2 ! > 9 6 : ;
U s e d b & T . .
@
$ingle antenna +port 0, Open7loop transmit diversit& Open7loop spatial m(ltipleing 'losed7loop spatial m(ltipleing MU7MIMO 'losed7loop transmit diversit& $ingle antenna +port9, 5daptive single7stream and d(al7stream beamforming 5daptive single7 stream d(al7stream and >7stream beamforming
Applica&le Scena!i $ingle7antenna transmission $(itable for cell edge here the channel condition is comple and interference is large or high7mobilit& or lo $=? sit(ations
,u!!ent eNdeB es es
$(itable for high U) mobilit& and comple reflection environment
es
$(itable for good channel condition #rovides high data transmission rate
es （ -.. ）
$(itable for to orthogonal U)s Used to increase cell capacit&
es
$(itable for cell edge lo mobilit& and lo $I=?
es （ -.. ）
$(itable for cell edge to reject interference
es
$(itable for cell edge lo mobilit& and high $=?
es
5 ne mode in T)75 $(pports a maim(m of eight la&ers Increases data transmission rate $(itable for lo mo bilit& and high $=?
=o
'oncepts
*!t
5 port is a logical port and does not necessaril& correspond to an antenna There can be m(ltiple ports The T) protocols s(pport a maim(m of eight ph&sical antennas #orts correspond to pilot formats hereas the n(mber of ph&sical antennas has not direct relationship ith the pilot formats
#ort 0 to port !% #orts for transmitting common pilots Us(all& the n(mber of ports for ph&sical broadcast channels and donlin8 control channels is the same as that for common pilots
#ort 9% 5 port defined in the T) for s(pporting single7stream beamforming The data of a single port can be eighted and mapped to m(ltiple ph&sical antennas
#ort 6% 5 port for locating the pilot
#ort : to port 1>% $imilar to port 9 $(pports a maim(m of ; la&ers The data of ; ports can be eighted and mapped to ; ph&sical antennas Used for d(al7stream beamforming
#ort 19 to port 22% '$I7?$ port
Maim(m n(mber of streams C =(mber of logical antenna ports D2 ports > ports or ; portsE
'oncepts
*ilts in the LTE system
'ell7specific reference signal +'?$,% '?$ is 8non as common pilot '?$ is (sed b& the control channels for channel estimation and demod(lation '?$ is (sed for demod(lation of TM1 to TM6 and ?$?F meas(rement
U)7specific reference signal at port 9% It is (sed for demod(lating TM:
.M ?$ at ports : to 1>% It is (sed for demod(lating TM; to TM@ and is the reference signal in ?@ and ?10 It s(pports MU7MIMO and demod(lation of a maim(m of eight la&ers
?eference signal at port 6% It is (sed for locating the U)
'hannel stat(s information meas(rement ?$ +'$I7?$,% It is (sed for meas(ring the channel *(alit& indication precoding matri indication and ?I '$I7?$ s(pports meas(rement of eight ports
$o(nding reference signal +$?$,% It is (sed for meas(ring the (plin8 channels and s(pports (plin8 sched(ling
Cell$%&ecifc 'e(e)ence Si!nal *C'S+ Normal
CP
downlink reference signal map relationship.
Open7oop Transmit .iversit&
In open7loop transmit diversit& +TM2, space7fre*(enc& bloc8 coding +$-B', is (sed if the n(mber of transmit antennas is 24 $-B' and fre*(enc& sitched transmit diversit& +-$T., are (sed if the n(mber of transmit antennas is >
$-B'% -or to7a& transmit +. 22 MIMO, the transmit diversit& (ses $-B' here A1 and 2 are the information to be transmitted before $-B' G indicates conj(gate operation f 1 and f 2 are different s(bcarriers and T 1 and T2 are different transmit antennas $-B' codes 1 and 2 to different antennas and s(bcarriers for transmission% 1 over T1 f 1 2 over T1 f 2 72G over T2 f 1 and 1G over T2 f 2 Therefore b& transmitting copies of 1 and 2 over different antennas and fre*(encies $-B' achieves diversit& gain
Open7oop Transmit .iversit&
S/B,0/ST-
-or >7a& transmit +. >2 MIMO or . >A> MIMO, $-B' and -$T. are (sed together In -$T. some of the transmit antennas are selected se*(entiall& in fre*(enc& for transmission
The transport format of $-B'H-$T. is as follos% 1 2 ! and > are information to be transmitted before coding4 f 1 to f > are different s(bcarriers4 T1 and T> are different transmit antennas4 G indicates conj(gate operation4 0 indicates no information transmitted In $-B'H-$T. 1 to > are coded to different antennas and s(bcarriers for transmission4 the transmit antennas are selected i8e $-B' $-B'H-$T. achieves diversit& gain b& transmitting copies over different antennas and fre*(encies
$patial M(ltipleing
$patial m(ltipleing means transmission of m(ltiple spatial data streams over different antennas in the same ?B The dimension of spatial channels is increased compared ith the single7antenna techni*(e Therefore spatial m(ltipleing increases s&stem capacit& and achieves spatial m(ltipleing gain $patial m(ltipleing incl(des to operations% la&er mapping and precoding .epending on hether the precoding matri is obtained based on the feedbac8 information of the U) spatial m(ltipleing is classified into open7loop spatial m(ltipleing +TM!, and closed7loop spatial m(ltipleing +TM>, The folloing fig(re shos the 22 spatial m(ltipleing
5daptive Mode 'onfig(ration
Mulit-Antenna transmit technologies can support dierent scenario transmit and mode. According to dierent scenarios, eNodeB support choose the most est MIM! mode.
Mode choice and switch four t"pe# O&en and clo%e loo& mode ada&tie choo%e and %witch O&en loo& ada&tie mode choo%e and %witch Clo%e loo& ada&tie mode choo%e and %witch -i mode choo%e
$% &'& MIM! and $% ('& MIM! support four mode choose and switch.
$% ()( MIM! onl" support open loop adapti*e mode choose and switch.
Conf!u)ation% o( 1I1O
Conf!u)ation o( 1U$1I1O
Application of MIM!
At persent, %+ +$$ can support " /&/&
1peci2cation of eNodeB
/&/0
Conf!u)at ion t3&e
MIMO
LBBPc
RRU3232
''U2/2"
3 × 10MHz
2 × 2 MIMO
1 LBBPc
2 (2T2R)
3
3 × 10MHz
4 × 2 MIMO
1 LBBPc
3
-
3 × 20MHz
2 × 2 MIMO
1 LBBPc
2 (2T2R)
3
3 × 20MHz
4 × 2 MIMO
3 LBBPc
3
-
'ontents
Bac"!und and O)e!)ie% f the LTE MIMO Techniques
*!inciples and Applicatin f the MIMO Techniques
*!inciples and Applicatin f Beamf!min"
#rinciples of Beamforming
Beamforming is a donlin8 m(lti7antenna techni*(e The transmitter of an e=odeB eights the data before transmission forming narro beams and aiming the energ& at the target (ser as shon in the folloing fig(re
Beamforming does not re*(ire the U) to feed bac8 information or (se m(ltiple antennas to transmit data The direction of incoming ave and the path loss information are obtained b& meas(ring the (plin8 received signal
The &enefits f &eamf!min" a!e as fll%s'
Increased $I=? in the direction of incoming ave from the U)
Increased s&stem capacit& and coverage
'lassification of Beamforming Techni*(es
-OA Beamf!min" and MIMO Beamf!min"'
.irection of 5rrival +.O5, beamforming% The e=odeB estimates the direction of arrival of the signal (ses the .O5 information to calc(late the transmit eight and targets the major lobe of the transmit beam at the best direction
MIMO beamforming% The e=odeB (ses the channel information t o calc(late the transmit eight forming a beam
In the indust!y$ the T-- system uses pen+lp Beamf!min" and the /-system uses clsed+lp Beamf!min". 1ua%ei eNdeB supp!ts pen+ lp Beamf!min".
'lassification of Beamforming +$ingle7$tream, $ingle7stream
beamforming means transmission of a single data stream in the same
O-.M reso(rce bloc8 It is s(itable for sit(ations of poor channel *(alit& $ingle7stream Ta8e
beamforming achieves diversit& gain b& 1 dB b& increasing the $=?
>7antenna as an eample The folloing fig(re shos single7stream
beamforming The data stream $ is eighted b& 1 to > and is sent to the fo(r antenna ports for transmission
'lassification of Beamforming +.(al7$tream, .(al7stream
beamforming means transmission of to data streams in the same O-.M
reso(rce bloc8 leading to spatial m(ltipleing It is s(itable for sit(ations of good channel *(alit& Ta8e
>7antenna as an eample The folloing fig(re shos d(al7stream beamforming There
are to data streams $1 and $24 each antenna has to eights i1 and i2 $1 is eighted b& fo(r eights% 11 to >14 $2 is eighted b& another fo(r eights 12 to >2 The eighted streams are s(mmed and sent to the fo(r antenna ports for transmission
)ngineering "(idelines of Beamforming
Before config(ring beamforming antennas &o( need to (nderstand the c orrespondence beteen the port =o and the co7polari3ation of cross7polari3ed antennas The folloing fig(re shos the connection beteen ??U ports and antenna element of the fo(r or eight antennas 5t present the ??U models in T) T.. that s(pport beamforming are ??U!2!2 ??U!2!! and ??U!2!9
>7antenna cross polari3ation mapping
>7antenna linear polari3ation mapping
>7antenna circ(lar polari3ation mapping
;7antenna cross polari3ation mapping
Beamforming 'ell 'onfig(ration
5dd an BB# b& r(nning the A-- BR- command ith Mde set to T--EN1AN,E
5fter adding the cell r(n the folloing commands to t(rn on the beamforming meas(rement sitch and algorithm sitch%
MO- MEAS3RES4IT,1' 3lintfMeasS%itch5S4BfN6alidMeas+ 78S4BfNRanMeas+78S4BfS!sMeas+79
MO- ,ELLAL:OS4IT,1' Lcal,ellId5;$ BfAl"S%itch5BfS%itch+79
S&ecifcation o( 4eam(o)min! Conf!u)ation T3&e MIMO
LBBPc
RRU3232
3 × 10MHz
4T4R Beamfom!"g
1 LBBPc
3
3 × 20MHz
4T4R Beamfom!"g
3 LBBPc
3
LBBP
RRU3232
Conf!u)ation T3&e MIMO 6 × 20MHz
4T4R Beamfom!"g 6 LBBPc
6
Conf!u)ation T3&e
MIMO
LBBP
RRU3233
3 × 20MHz
#T#R Beamfom!"g
3 LBBPc
3(eac$ RRU "ee% &'o f!e ）
#I of Beamforming Leadin"
2 Beamf!min" Enhanced the ,apacity Test Result in
Al%ays Leadin" in Beamf!min"
=:** R>
=:** R
=:** R7;
single7 stream beamformi ng
d(al7 stream beamformi ng 7st t supp!t
M(lti7User Beamformin g
7st t launch Sin"le+st!eam Beamf!min"
-ual+st!eam Beamf!min"
07?@
07;@
07?@ Lisilcon Balong:10 'hipset is the first to s(pport d(al7stream beamforming
Lisilcon Balong:00 'hipset is the first to s(pport single7stream beamforming
2;7717
2;7712
2;7217
J2Mbps
J>Mbps
J6Mbps
TM:
@190K
:!>0K
6010K
TM2
;2;0K
61@0K
9610K
#I of Beamforming Rele)ant
featu!es $ingle7stream beamforming m(st be enabled before d(al7stream beamforming Influence
n the C*I $ingle7stream or d(al7stream beamforming has the folloing infl(ence on the #I% 'ell average thro(ghp(t If the single7stream and d(al7stream beamforming is enabled the signal energ& received b& the U) is increased the M'$ is increased at the same U) position beamforming achieves higher cell average thro(ghp(t than transmit diversit& In comparison ith no beamforming single7stream beamforming increases the cell average thro(ghp(t b& 19K to 29K In comparison ith single7stream beamforming adaptive single7stream and d(al7stream beamforming increases the cell average thro(ghp(t b& more than 10K
Beamforming compared ith 2? diversit& +U, <
!0K gain in cell average thro(ghp(t
<
90K gain in cell edge (ser thro(ghp(t
Beamforming compared ith 22 MIMO +., <
19K gain in cell average thro(ghp(t
2!K@0K increasing in edge < >0K gain in cell(ser edgethro(ghp(t (ser thro(ghp(t
5daptive MIMO and Beamforming
Nith adaptive beamforming and MIMO the U) ala&s (ses TM of high spectral efficienc& (nder the same channel condition In comparison ith non7adaptive MIMO or beamforming adaptive MIMO and beamforming significantl& increases average cell thro(ghp(t If beamforming is (sed d(e to the overhead of U)7specific reference signal the n(mber of reso(rce bloc8s is red(ced Therefore in case of good channel *(alit& beamforming thro(ghp(t is slightl& loer than MIMO thro(ghp(t 5t high U) mobilit& +higher than 120 8m/h, the e=odeB cannot trac8 the channel change acc(ratel& according to the so(nding reference signal In this sit(ation beamforming is not s(itable 5daptive beamforming and MIMO +lo 5daptive beamforming and MIMO +high mobilit&, mobilit&,
5daptive MIMO and Beamforming
The B/MIMOA-A*TI6ES4IT,1 parameter is (sed to select adaptive beamforming or M IMO The e=odeB selects beamforming or MIMO according to the val(e of the parameter the U) movement speed and $I=?
If the val(e of the parameter is NOA-A*TI6E the e=odeB does not s(pport adaptive Beamforming and MIMO
If the val(e of the parameter is T-B/A-A*TI6E the e=odeB s(pports adaptive TM2 +transmit diversit&, and beamforming There are to scenarios% lo U) mobilit& and high U) mobilit& o U) mobilit&% -or U)s that do not s(pport ?@ single7stream beamforming +TM:, is (sed4 for U)s that s(pport ?@ single7stream beamforming +TM: or TM;, is (sed at lo $I=? and d(al7stream beamforming +TM;, is (sed at high $I=? Ligh U) mobilit&% Transmit diversit& is (sed
If the val(e of the parameter is MIMOB/A-A*TI6E the e=odeB s(pports adaptive transmit diversit& d(al7stream MIMO +TM!, and beamforming There are to scenarios% lo U ) mobilit& and high U) mobilit& o U) mobilit&% -or U)s that do not s(pport ?@ single7stream beamforming +TM:, is (sed at lo $I=? and d(al7stream MIMO +TM!, is (sed at high $I=?4 for U)s that s(pport ?@ single7stream beamforming is (sed at lo $I=? and d(al7stream beamforming +TM;, is (sed at high $I=? Ligh U) m obilit&% Transmit diversit& is (sed at lo $I=? and d(al7stream MIMO +TM!, is (sed at high $I=?
'omparison Beteen Beamforming and Other Techni*(es
Tho(gh a space diversit& s&stem or intelligent antenna s&stem has m(ltiple transmit or receive antennas the& can transmit onl& single7stream data 5 MIMO s&stem can transmit single stream or m(ltiple streams depending on the channel *(alit&
MIMO re*(ires that the n(mber of receive antennas is not less than the n(mber of transmit antennas $pace diversit& and intelligent antennas do not have this re*(irement
