| Full Text |
MJiTHOD iVNi3 AI'PARATUS FOR ALLOCATING RESOraCES IN A MOLTIPLE-INPtTr rJLTIIlE OOTPUT (MIMO) COWVIXTMICATION SYSTEM
BACKGROUND
Field
11001] I'he preiient invention relates generally to data communication, and more j;]>eci{i';ally i:C' techniques for allocating downlink resources in a multiple-input multiple-ijutput (MINEC)) communication system.
;3ack]j round
11002J W:ir;5less communication :jyj;tems arc widely deployed to provide various i)^es Cff communication such as voice, data, and so on, for a number of users. These systems ma} lie based on code division multiple access (CDMA), time division multiple access (TDM\), frequency division multiple access (FDMA), or some other multiple access; techniques.
11003j A, multiple-input multiple-output (MIMO) communication system employs JtLultipJe (Ni) transmit antsnnas and multiple (NK) receive jintennas for transmission of multiple ind2]:>e:ndent data streams. In one conmion MIMC' system implementation, of the data streamij are transmitted to a single terminal at any given time all. However, a jriultiple access coromunication system having a base station with multiple antennas may aho coiicurrendy communicate with a number of terminals. In tiiis case, the base station employs a number of antemias and each terminal employs NR antennas to leceive one or more of the multiple data streams.
11,004] The connectio:^ between a multiple-antenna base station and a single multiple-anti^rma teiminal is called a MIMO channel. A MIMO channel formed by these MT trari;i;mit and NR receive anr,ennas may be decomposed into Nc independent channels, with Nc ^ min {NT. NR}. Each of the Nc i.n
j-ansrnissiort capacity) if the additioniil ^iimensionalities of liiese subchannels created by
Jie multiple crartsmit and receive antennas are utilized.
[10051 Ivcich MIMO channel betw^^n the base slalicm and a terminal typically
iKpehmces different link characteris-tics and is ass(x:iated with different transmission
:apabi ity, so the spatial subchannels available to each temiinaJ have different effective
;apacaies. lEfiicient use of the available downlink re.sources (and higher throughput)
nay hi achieved if the NQ avgdlable spatial subchannels are effectively allocated such
iiat data is tri:tr.smitted on these subchiarinels to a "propeir*' set of terminals in the MIMO
iystern.
1006] lliers is therefore a need in the ait for techrtiques to allocate downlink
;-
SUMMARY
"1007] /liipects of the invention provide techniques to increase the downlink >erfonaanc€: of a wireless communication system. In an aspect, data may- be lansmitted fio3 a base station to one or more terminds using one of a number of diffennt optnating modes. In a MIMO mode, all available downlink data streams are illocatjd to a single terminal that employs multiple antennas (i.e., a MIMO terminal). ji an N-SIMO mode, a single data stream is allocated to each of a number of distinct erminiils, with each terminal employing multiple antennas (i.e., SIMO terminals). And :n a niJted-niD-de, the dov/nlink resources may be allocated to a combination of SIMO itid MMO '.errtiinals, with both type^. of terminals being simultaneously supported. By 'lansmitting data simultaneously to multiple SIMO terminals, one or more MIMO ierminjds, or a combination therecf, the tmnsmission capacity of the system is increased.
110081 ii iDOtlier aspect, scheduling schemes are provided to schedule data iransmissions to aciive tenninals. A scheduler selects thie best operating mode to use based on var:.ous factors such as, for example, the services being requested by the lermiriiils. Ii addition, the scheduler <:an perform an adiiiljonal level of optimization by .selecting a particuhir set terminals fcr simultaneous data transmission and assigning ihe available u-ansmit antennas to the selected tennin such that high system jxsrfornance iu:d o requirements are achieved. several scheduling schemes juiteniiii assijumient provided described below.>
[1009] A specilic emtl^cdiment o:" the invention provides a method for scheduling iownlink djla trarisinission to a number of lerminuls in a wireless conimunication system. In accordaince with the method, one or more sets of terminals arc formed for [)ossil): e dat;a ti ansniission, witli each sei: including a unique combination of one or more :ndinj;; to si)ecific assignments of a number of transmit antennas to the one or iiore lermiii£.lii in the h>pothesis. The performance oi'
10101 I;3.cti transmit antenna may be used to trans:iiit an independent data stream. To achieve hij^x peiformajice, each data stream may be coded and modulated based on a :ichemi; seU;cttimii'e for the antenna u*;ed to transmit the data stream.
10111 Tenriinals desiring data transmission (i.e., "active" terminals) may be jrioritdzed ta.5<:d on various metrics and factors. the priority of active terminals nay then be used to select which tenmnal considered for scheduling iiijsign a able transmit antennas selected terminjils.>
11012] llie in^^ention further provides methods, systems, and apparatus that impleiiient various aspects, embodiments, and features of the invention, as described in lurther detail below.
BRIEF DESCRIFCION OF THE DRAWINGS
11013] Ihe features, nature, and advantages of the present invention will become more apparen: irronj the detailed descdption set forth belc^w when taken in conjunction ^vith the dr^.\i'ingii in which like reference characters identify correspondingly thLTOUghout and >vherein:
11014] FIG. 1 is a diagram of a multiple-input multiple-output (MIMO) c:omrat.nicat^cin system that may be designed and operated to implement various aspects mid ernbodirriiJiits of the invention;
[1015] FIG. 2 is a ilow diagram of a process to schedule tenninals for data
b-ansrtissioii, in accordance with an onbodiment of the invention;
[lOl^l FIG. 3 is a flow diagram of a process to assign transmit antennas to receive
anten::ias usinj; a "^max-max*' criterion, in accordance with an embodiment of the
invention;
[1017] ¥IG, 4 is a flow diagram for a priority-based scheduling scheme whereby a
iiet of one or :ir.'Dre liighest. priority terminals is consi(lerl:^d for scheduling, in accordance
^nth an embodiment of the invention;
1018] ¥iG. 5 hi a block diagram of a base station and a number of terminals in the
AIIMC comcDu:nicadon system;
11019] ■ FIG, 6 is a block diagram of an embodiment of the transmit portion of a biise
sv-ation capable of processing data for transmission to the terminals based on the
iivailatleCSI;
1.11020] IrlG.. 7 is a block diagrara of an embodiment of the receive portion of a
lerminid;
110211 FIGli. 8A and 8B are block diagrams of an embodiment of a channel
]vIIMO/data processor and an interference canceller, respectively, of a receive (RX)
IvDMO/data paicessor at the terminal; and
11022] FIG. 9 shows the average throughput for a MEMO communication system
^vith fC'Ur tninsmit antennas (i.e., NT = 4) and four rec-eive antennas at each terminal
(i.e., N
DETAILED DESCRIFIION
[1023] FIG. 1 is a diagram of a multiple-input multiple-output (MIMO) communicaten system ICO that may be designed and operated to implement various cspects and embodiments of die invention. MEMO system 100 employs multiple (NT) ininsnriit anti5r]ji.2;s and multiple (NR) receive antennas for data transmission. MIMO sj-stem 100 is effectively formed for a multiple access coroinunication system having a c-ase stiition QiS) 104 that can concuiTcntly communicate >vith a number of terminals [T) 106. In tliis case, base station 104 employs multiple antennas and represents tlie ::iultipli5-input (MI) for downlink transmissions from the base station to the terminals. [1024] A siet of one or more "conmunicating" terminals 106 collectively represents iJi5 multiph;-output (MO) for downlink transmissions. As used herein, a
comiiiunicaung lenmnai is one mat receives user-speciT:ic aaia rrora uie case siauon, ind an "activ'«,;" tenninal is one that desires data transiTiission in an upcoming or future lansniissior interval. Active terrainjds may include terminals that are currently i;omir:unicatirg.
11025] Ml^'IO system 100 may be designed to implement any number of standairds itnid designs ibr CDMA, 'TDMA, FDMA, and other multiple access techniques. The CDNLfi. standards include tlie IS-95, (xima2000, and W-CDMA standards, and the TDNL^. standards includt; the Global System for Mobile Communications (GSM) :;l:and£ud. Thtse standards are known in the art and inco:rporated herein by reference. [1026] ISCIMO system. 100 may b
I1027J For the example shown in FIG. 1, base station 104 concurrently i:ommi:nicat;j> v/ith termnials 106a tlux»ugh 106d (as indicated by the solid lines) via multiple antiiuruis available at the base station and multiple antennas available at each lermiriiJ. Tcnrlaals 106e through lOfh may receive pilot references and other signaling irioraxition from base station 104 (as indicated by the dashed lines), but are not itceiving us
11028] Eadi terminal 106 in MIMO system 100 employs NR antennas for reception o:: one or more data streams. Generally, the number of antennas at each terminal is ujual to or greater than the number of data streams transmitted by the base station. However, the tJJttninals in the system need not all be equipped with equal number of i€ceiv2 antennas.
11029] FQ:r.MIMO system 100, the number of antentiai; at each of the terminals (NR) is typically g::iiiU:er than or equal to the number of antennas at the base station (NT). In this caie, for the downlini:, the numb€;r of spatial subchannels is limited by the number of trariimit anxnnas at the base station. Each transmit anienna may be used to send an
iadependeni: data stream that may be coded and modulated based on a scheme supported by tht spatial !;ubcharmel associated with the MIMO channel between the base station and tli'j selei;i.e(J terminal
[1030] i\ ;{3ejcts of the invention provide lecl-miques to increase the performance of a v;irele>s communication system. These techniques may be advantageously used to increase the ii
[1032] la another aspect, scheduling schemes are provided to schedule data transirissions to active terminals. A scheduler selects tlie best operating mode to use based on v.;irious factors such as, for example, the services being requested by the ujrrmnals. In addition, the scheduler can perform an additional level of optimization by s^lcctiig a i^jorlicular set of terminals for simultaneous dati transmission and assigning tlie a.\ailable transmit antennas to die selected terminals such that high system perfonnance and otiier requirements are achieved. Several scheduling schemes and antenna assignment schemes are described in further detail [)eIow. [1033] ^Vith MMO, multiple independent data streiuiis may be transmitted from the base siation via multiple transmit antennas to one or mote scheduled terminals. If the propagation environment has sufficient scattering,. MIMO receiver processing :echn];([ues riiiy be used at the terminals to efficiently exploit the spatial dimeriiiionalities of the IvUMO chaiinel to increase triinsmission capacit}'. MIMO ::^K:eivc:r procisiJJting techniques may be used when the bai;e station is communicating Alth multiple: ajrminals simultaneously. From the terminal's perspective, the same
r5ceiv(5r prooi^ssing techniques may be used to process Nj different signals intended for rliat termins.l (i^.g., a single MIMO terminal) or just one of the Nr signals (i.e., SIMO terminals).
[1034] A.:; shovm in HG. 1, the tenninais may he i-andomly distributetl in the base station's coverage jjrea (or "cell") or may be co-located-. For a wireless communication systeir., the iialc ch;iracteristics typically vary over time due to a number of factors such as fading am:! ]:nult:path. At a particular instant in time, tl:,e channel response between ilie base statjoii's aiTay of Kj transmit antennas and the NR receive antennas for a single terminal ma]/ be characterized by a matrix H v/hose elements are composed of independent (j;iijssian random variables, as follows:
tf/here; H is 1J)S channel rei^onse matrix for the terminal, £jid hij is the coupling between tlie baie station's f-th transmit antenna and the tcrminal'sy-th receive antenna. [1035] i'u. shov^-n in equation (1), the channel estimates for each terminal may be r»;prest:nted v-'idi a matrix, having N^-xN^ elements corresponding to the number of
ti'ansrnit antennas at the base station and the number of receive antennas at the tenninal. Elach clement of the matrix H describes the response for a respective transmit-receive antenna paL* between the base station and one terminal. For simplicity, equation (1) jescril'cs a channel characterization based on a flat fading charmel model (i.e., one ■:omp"l
1036] The active terminals in the MEMO system periodically estimate the channel :-cspon5e for each transmit-receive antenna pair. The channel estimates may be "acilitated in a number of ways such as, for example, with the use of pilot and/or data ilecisicn directed techniques known in tlie art. The channel estimates may comprise the complex-value channel response estjfnate for each triinsmit-receive anteima pair, as
jescribed ahovij in equation (1). The channel estimateis give infomiation on the 3*ansrnissiori characteristios of each of the spatial sub:;hafinels, i.e., what data rate is suppC'itable on each subchannel with a given set oi' li^uismission parameters. The information jpven by tlie channel estimates may be distilled into a post-processed 5tgna]-to-noise ]?lu5.-interfsrence ratio (SNR) estimate for each spatial subchannel (described bo!x^^v), i^r some other statistic that allows the Ixajismitter to select the proper tiansrnissioii piiranaeters for that spatial subchannel. Typically, this process of Jerivation of the essential statistic reduces the amount of d£ta required to characterize a :hannel. I:Q eittier cai;e, this infonnation represents one form of channel state iuforraation [CSI) that may be repon:ed to the base station. Other forms of CSI may ilso be repoite«J and are described below.
[1037] The aggregate CSI received from the collection of terminals may be used to [1) selsct a ""best" set of one or more terminals for data transmission, (2) assign the ivailaWe transrrdt antennas to the sekx:ted terminals in the set, and (3) select the proper :Dding and modulation scheme for each transmit antenna. With the available CSI, /arioui; scheduling schemes may be designed to maximize tlie downlink performance by ivaluiiing vv'Iiich specific combination of tenninals and antenna assignm^ints provide he best syitsin performance (e.g., the highest throughput) subject to any system oonstrzints and requirements. By exploiting the spatial (and possibly frequency) 'signaiures" of the individual active terminals (i.e., their channel estimates), the average down'tink thxiughput can l)e increased.
;i0381 Ttic terminals may be scheduled for data trarsmission based on various "actors. One -set of factors may relate to system constraints imd requirements such as the iesirec. quality of service (QoS), max:imum latency, average data rate, and so on. Some or all of the.-!!5 factors may need to te satisfied on a per terminal basis (i.e., for each 'erminid) in u multiple access commuiiication system. Another set of factors may relate 10 system pc:rrain3ance, which may b
110391 The scheduling schemes can be designed to select the best set of terminals for si:3iultaneous data transmission on the available tiransnaission channels such that iiystem perfomiance is maximized while conforming to the system constraints and resquirsmentii. Por simplicity, various aspects of the invention are described below for a
jvtnvKJ sysi.em without OFDM in which one indeptMident data stream may be iransxnitted by the base station fron} each transmit aji:eiina. In this case, (up to) NT •ndep NT-
110401 for simplicity, the number of receive antennas in assumed to be equal to the numtHi: of ti aa:;:rait antennas (i.e., NR = NT) for much of thci description below. This is not a nscessaiy coniiition since all of the analysis applies for the case where NR ^ NT-11041] S::heduling of data transn:isi;ion on the dov/nhnk comprises two parts: (1) iielection of ont or more sets of tenninals for evaluation, and (2) assignment of the available trajHrtdt fintennas to the teiminals in each set. All or only a subset of the active terminals; may be considere»:heduler to optimize ])eiformance by exploiting muhi-user diversity snvirorment.
[1042] Li order to determine the "optimum" transmission to a set of terminals, the DNRs or soric other sufficient statistics are provided for each terminal and each spatial :ubcha'ineL If die statistic is the SNR. then for each set of t-^rminals to be evaluated for :iita transmiiisicii in an upcoming transmission interval, a h^^othesis matrix T of "post-C'locessed'* SMls (defined below) for this terminal set may be expressed as:
7/here y^j is trie post-prc»cessed SNR for a data stream (hypothetically) transmitted ;'-om til 3 /~th uansmit antenna to the;-th teraiinal.
;1043] In tlis; N-SIMC> mode, the Nr rows in the h>'pot]iesis matrix T correspond to >rT vectors of SNl^s from NT different temoiiials. m t!iis mode, each row in the lypotbesis riiiitrix T gives tlic SNR of each transmit data stream for one tenninal. And II the nixed-ai(xle. for a particular MMO terminal designated to receive two or more data stream:;, t.hat 'terrain.il*s vector of SNRs may be replicated such that the vector ippears in a:> Tiany rows as the number of data streams to be transmitted for the terminal ;i.e., oie vo'M pe uswl for cach SIMO or MIMO terminal, and the scheduler may be designed to mark ind ev aluatc: l:hi:se different types of tiirminals accordingly.
;;1044] i^i.1 e^ach terminal in the set to be evalaatcd, the NT (hypothetically) lansrnitted -lata stn^aras are received by the terminal's NR jeceive antennas, and the NR :Tt:ceiv(d signals can be processed using spatial or spa^e-time equalization to separate )\xt the NT trginsniitted data streams, as described below. The SNR of a post-processed -lata st-eam (i.e., after equalization) may be estimated atic comprises the post-processed SNR for that data stream. For each tenninal, a set of N'T post-processed SNRs may^be jrovidid foi the NT data streams that may be received by that terminal. 1045] ir a successive equalization and interference cancellation (or "successive cancellation") receiver processing technique is used at a tenninal to process the received .signals, ther iht post-proc:essed SNR achieved at the tenninal for each transmitted data :;iream depends, on the order in which the transmitted data streams arc detected (i.e., 'l^modulatecl and decoded) to recover tlie transmitted data, as described below. In this i:;ase, a nurcber of sets of SNRs may be provided for each terminal for a number of jjDSsibie deta»:;tion orderings. Multiple hypothesis matrices may then be formed and (jvaluuled to detennine which specific combination of terminals and detection ordering jjrovidi^s the test system performance.
[1046] In any case, each hypothesis matrix T includes the post-processed SNRs for ;i specific set a:" tenoinals (i.e., hypoth»esis) to be evaluated. These post-processed SNRs represent the^ !5:S[Rs achievable by the teiminals and are u&^d to evaluate the hypothesis. 110471 FTG. 2 is a flow diagram of a process 200 to schedule terminals for data iransmission, in accordance with an embodiment of the invention. For clarity, the overall proceiis is first described and Ae details for some of the steps in the process are Inscribed subi^icjuently.
[1048] Iri:.ti2tlly, metrics, to be used to select the best set of tenninals for data transrcissioij arc imtialized, at step 212. Various pexfonn^Lnce metrics may be used to 3valu;i:e the teimnal sets and some of these are descrit>ed in further detail below. For example, a ];erfbrmance metric that miiximizes system tiir^tighput may be used. [1049] A. (rigw) set of one or moie active terminals is tien selected from among all active termin:aj.s co:asider«;d for scheduling, at step 214. This set of terminals forms a hypothesis to- IK' evaluated. Various t
all terriinalJ in i:he hypothesis form the hypothesis matrix P shown in equation (2). [1050] Fcir isach hypothesis matrix F of Nx transmit antennas and Nt tenninals, there iire Nr factorial possible combinations of assig:iments of transmit antennas, to terminals (.e., N7! sub-hypotheses). Thus, a particular (new) combination of-antenjia/teriminal assignments is selected for evaluation, at step 218. This particular :omb;inatior of ;antenna/termina] assi|pments forms a sub-hypothesis to be evaluated. [1051] 1 he sub-hypothesis is then evaluated and the performance metric (e.g., the jystern throujjliput) corresponding to this sub-hypothesis is; determined (e.g., based on Jie SNRs fcr the sub-hypothesis), at step 220. This peifonnance metric is then used to update the perfc-rmance metric corresponding to the current best sub-hypothesis, at step 222. Specifically, if the performance metric for this sub-hypothesis is better than that of he cicTent biii;t sub-hypothesis, tfien tliis sub-hypothesis becomes the new best sub-lypothesis, and the performance metric and other terminal ruetrics corresponding to this 5ub-hypothe.>is arc saved. The perfonnance and terminal metrics arc described below. '10521 A detennination is then inade whether or not all sub-hypotheses for the t: jrreril hypcn:liesis have been evaluated, at step 224. If £J1 sub-hypotheses have not been .;valujij:ed, the process returns to step 218 and a different and not yet evaluated combination of antenna/t-srminal assignments is selected for evaluation. Steps 218 ilirouj^i 224 an:: repeated for each sub-hypothesis to be evaluated [1053] I;' all sub-hypoiJieses for a particular hypothesis have been evaluated, at step :224, a detemination is then made whetlier or not all hjpotiieses have been considered, at step 226. If ail hypotheses have not been considered, tlieu the process returns to step
214 aid a cii:iT:5renl: and not yet considered set of teniiiniils is selected for evaluation. Steps 214 throujjh 226 are repeated for each hypothesis to bs considered, [1054] ff all hypotheses have been considered at step 226. then the specific set of terminils scheduled for data transmission in tJie upcoming transmission interval and their a:jsign€:c t:rans:Enit antennas arc kaow^n. The post-processed SNRs corresponding to Jiis se: of terniinaiii and cintenna assignments may be used, to select the proper coding and nixiuladoji scJiemes for the data streams to be trginsinitted to the terminals. The scheduled tiitiiijmis^uon interval, antenna assignments, coding and modulation schemes, 3ther :nfon::i.;:ition, or any combinaton thereof, may be conveyed to the scheduled tcjrminals (e.fi;., via a control channel), at step 228. /Jternatively, the teiminals may perfonn "blind" detection and attempt, to detect all trans:mi.tt3d data streams to determine which ones, if J=t3iy, of the ilata streams are intended for them..
[1055] If tl:ie scheduling scheme requires other system and terminal metrics to be maintained {j;.g. the average data rate over the past K transmission intervals, latency for lata triinsmisision, and so on), then tht^se metrics are updated, at step 230. The terminal aietria; ma}- \y^> usc:d to evaluate the performance of the individual terminals, and are described bdov;. The scheduling is typically performed for each transmission interval. ;;10561 F 0 r a given hypothesis matrix F, the scheduler evaluates various oombijiationi; of transmit antenna and terminal pairings (i.e., sub-hypotheses) to 'istermine ths t^est assignments for the hypothesis. Various assignment schemes may be ised tc assi|p: transmit antennas to the terminals to achieve various system goals such as :iiimes5, ma
1057] ti one antenna assignment scheme, all possible sub-hypotheses are evaluated '>ased on a particular performance metric and the sub-hypothesis with the best
l)3rfori3ance :ii]i:tric is selected. For e£x:h hypothesis matrix F, there are Nj factorial li.e., TJT!) :?i)ssible sub-hypotheses that may be evaluated. Each sub-hypothesis corresponds i;o a specific assignment of each transmit antenna to a respective terminal. Bach sub-h}i30lJiesis may thus be re])resented with a vector of post-processed SNRs, which nay be expressed aii:
vdiere ;/, ^ is t:h(5 post-processed SNR for the f-tti traiisniil EJitenna to they-ih tenninal,
und the subi^cripts {a, by ... and r) identify the specific terminals in the transmit imtenna/termiria;, pairings for the sub-hypothesis.
11058] Eich isub-hypothesis is furtlier associated with a performance metric, Rsub-hyp, '?/liich roay be a function of various ftictors. For example, a performance metric based ;)n the post-prccesscd SNRs may be exprsssed as:
■ "'here /(*) is a particular' positive real function of the argument(s) within the
parenth'jsis.
I.L059] Vaiioijs functions may be used to formulate the performance metric. In one oaibodinent, a f'j;nction of the achievable throughput for all NT transmit antennas for the >iib-hyf othesis rnay be used, which may l:>e expressed as:
Mhere .",• is tJie throughput associated with the i-lh transmit antenna in the sub-liypothcsis, aid may be expressed as:
rvhere ct is :?. positive constant that reflects the fraction of the theoretical capacity I oliieved by the coding anc modulation scheme selected for the data stream transmitted 01 the t th transrdt antenna, and yt is the post-processed SNR for the i-th data stream, []J)60] Ths first antenna assignmeiit scheme shown in FIG. 2 and described above ] £:l)reser.ts a sp3^:;ific scheme that evaluates all possible combinations of assignments of transmit antennae; to terminals. The total number of potential sub-hypotheses to be ev,aluated by 4it schsduler for each h>pothesis is NT!, which may be large considering fat a lirge number of hypotheses may need to be evaluated. The first scheduling scheme perfortiis an exhaustive search to determine the sab-hypothesis that provides the "opdmal" system performance, as quantified by the perlbnnance metric used to select the best sub-hypothe«;is.
[1061] A number of techniques may be used to ie
[1062.] In 11 se<:ond iuitenna assignment scheme a maximum-maximum critenoi is used to assign trans:mit antennas tlie terminals in the hypothesis being using t max-max criterion each tra:asmit antenna assigned partici.lar tejminal that achieves be snr for transmit antenna- assigiment : lurrforraed one at time.>
[1063] f'lG'. 3 is a flow diagrarn of a process 300 to assign transmit antennas to terminals using the max-max criterion, in accordance v/ith an embodiment of the invention. Thii pr(x:essing shown in FIG. 3 is peribrmed for a particulai* hypothesis, v^hich corres]K:nds to a s]3ecific set of one or more terminals, InitiaUy, the maximum post-p:"Oces5ed SKR in the hypothesis matrix T is detem:ined, at step 312. This maxiirum 5I>JR corresponds to a sf»ecific transmit antenna/terminal pairing, and the transmit anteimia is assigned to this terminal, at step 314. This transmit antenna and terminal are then removed from the matrix T, and the maixix is reduced to dimension (N^. - l)x(ri-p -1) by removing both the column corretponding to the transmit antenna
and the row carresponding to the teradnal just assigned, at step 316.
[1064] /it step ;?18, a determination is made whether or not all transmit antennas in
Jie h)i)0the^.;5 have been assigned. If all transmit antennas have been assigned, then the
antenna assigaments are provided, at step 320, and the process terminates. Otherwise,
die p::
Planner.
[1065] Ones: the antenna assignmisnts have been made far a given hypothesis matrix
r, the perfoniiance metric (e.g., the system throughput) ccnesponding to this hypothesis
:iiay he dete:;:nx;ned (e.g., based on the SNRs corresponding to the antenna assignments),
ii5 shown in equations (3) and (4). This performance metric is updated for each
hypothesis, '^'hen all hypotheses have been evaluated, thie best set of terminals and
antenna assigixrnents are selected for data transmission in the upcoming ixansmission
interval.
[IGiiS] Table 1 shows an example matrix F of post-processed SNRs derived by temdnals :.n a 4x4 MMO systera m which the base station includes foux transmit antennas imc] each terminal includes four receive Entennas. For the antenna assignment scheme biased on the rnax-max criterion, the t>est SNR (16 d.B) in the original matrix is achieved by transmit antenna 3 and is assigned to tennin;il 1, as indicated by the shaded box in thv. itird row of the fourth cohimn in tlie table. Transmit antenna 3 and terminal 1 ar-j ther removed frora the matrix.. The best SNR (14 dB) in die reduced 3x3 matrix is achieved t>)' ho\h trarismit antennas 1 and 4, which are respectively assigned to tcnrinals 'i txid 2. The remaining transmit antenna 2 is then assigned to terminal 4.
[1067] Table 2 sho>vs the antenna assignments using the max-max criterion for the exanple matrix F shown in Table I. For temiinal 1, the best SNR (16 dB) is achieved whei processing the signal transmitted from transmit antenna 3. The best transmit antijinas fcr the other terminals arc also indicated in Table 2. The scheduler can use this infonniition to select the proper coding and modulation scheme to employ for data trar-j'tnissioa.
|;10681 Hie ycheduling scheme described in FIGS, 2 and 3 represents a specific ;;cheiTi(; that ir/iilu£ites various hypotheses corresponding to various possible sets of .n:tive :enm:ii:Ll:; desiring data transmisston in the upcoming transmission interval. The U)tal number of hypotheses to be evaluated by the scheduler can be quite large, even for ;i small numtKjr of active terminals. In fact, the total numte: of h>pothcses, Nbyp, can be (jApresi^das:
A'here Nu is tdc number of active terminals to be corsidered for scheduling; For cxarap! e, if 'Nu == 8 and NT = 4, then Nhyp = 70. An exhaustive search may be used to dstennine t::.<: hypothesis the particular anterina assignments that jjrovidcs optimal system performance as quantified by metric used selwa be:-l and antenna assignments.>
11069J 0 ih NT.
[1072] I'IG. 4 is a flow diagram for a priority-based scheduling scheme 400 wher&by a 3e: of NT highest priority terminals is considered for scheduling, in accordance v/ith an embodiment of the invention. At each frame interval, the scheduler 3xamj.nes th;! priority for iill active teintninals in the list and selects the set of Nj highest ;)riorir/ tenrixiiiiJs, at step 412. The remaining terminals in the list are not considered for ^i;hedLiling. 'JIIK; channel estimates fc-r each selected tenniral are then retrieved, at step 414. F'or example, the post-processed SNRs for the selecis
;i073] l?.,e 'Sr transmit antennas are then assigned to Ac selected terminals based 3n the chaiuiel esimatej; and using jmy one of a number of antenna assignment Tichenruss, at Mep 416. For example, the antenna assigmnents may be based on an idiaustive ;ic:an:h or the max-max criterion described above. In another antenna issigTiment scheme, the transmit antennas are assigned to the terminals such that their ^riorilJes are; iionnalized as close as pos^iible, after the termi;tial metrics are updated. "1074] The data rates imd coding iind modulation schemes for the terminals are then determined bt^cd on the antenna assij^nments, at step 418. The scheduled transmission iitcrval and liiita rates may be repo::ted to the scheduled terminals. Hie metrics of ;icheduled (and unscheduled) terminals in the list are up«3ated to reflect the scheduled data tnmsmiiyion (md non-transnrission), and system metrics are also updated, at step 420.
|'10751 \'Brious metrics and factors may be used to determine the priority of the active erminalj:.. In an embodiment, a "score" may be maintained for each terminal in ihe list and f«:>r each metric to be used for scheduling. In one embodiment, a score indicat.ve of ai average throughput over a particulai" averaging time interval is inaintaned for each active terminal. In one implementation, the score ^„(k) for
lermirud n at frame k is computed as a linear average throughput achieved over some time interval, md can be expressed as:
where r^(z) is tie realized data rate (in unit of bits/frame) fox teiminai n at frame i and may be comjuititd as shown in equation (4). Typically. r^(i) is bounded by a particular
aiaxiir.um a:hievable daUt rate, rmax. iind a particular niimmum data rate (e.g., zero). In another impIcrnentation, the score p^{k) for terminal n m frame fc is an exponential avera.gs thrc iighput achieved over sorae time interval, and can be expressed as:
vi^here a is a dme constant for the exponential averaging, with a larger value for a :orresii>onding \o a longer averaging tjne interval
[1076] ^^^hinl a terminal desires data transmission, it is added to the list and its score if. initialized to zen). The score for each temainal in the list is subsequently updated on 3ach Irame. 'vSHiencver a terminal is not scheduled for tritnsmission in a frame, its data rate.fa: the iiTinie is set to zero (i.e., r^(/:) = 0) and its score is updated accordingly. If a
frame is received in error by a terminal, the terminal's effective data rate for that frame may bt; set to zero. The frame error may not be known iimnediately (e.g., due to round nip delay of rn acknowledgment/negative acknowledgnent (Ack/Nak) scheme used for Jie da:!a traniaris&ion) but the score c;m be adjusted accordingly once this information is ivailable.
[1077] Trie pricrity for the active terminals may also be determined based in part on jystern conso^nts and requirements.. For example, if tlie maximum latency for a ^axticijlar tcnninal exceeds a threshold value, then the terminal may be elevated to a ligh piiorit}.
1078] Othsr factors may also be considered in detemdning the priority of the active :ermin.ils. Ox such factor may be related to the type of data to be transmitted to the *ermin.ils. E^elay sensitive data may be associated with higher priority, and delay insensitive rruiv be associated with lower priority. Retransmitted data due to decoding errors for a prior transmission may also be associated witli higher priority since other ;>rocesi;es may be awaiting the retransimitted data. Another factor may be related to the ij^e of datii service being provided for the terminals. Other factors may aJso be i;onsid
value;; (i.e., Iu|2h or low priority, depending on whetiicr or not the constraints and ::^Mquiiement> have been violated) and the scores represent "soft" values. For this i;:mbodinienl, terminals for v/hich the system constraints and. requirements have not been :iiet ai-t; imir^jdiately considered, alonj? v/ith other tenninals based on tlxeir scores. i'lOSOJ Pi. piiority-bas-^d scheduling scheme may be designed to achieve equal average thrctughput (i.e., equal QoS) for all tenninals in the list. In this case, active venniniils ars prioritized based on thieir achieved average throughput, which may be ticterroined .u; ;ii]iown in equation (6) or (7). In this ptioiity-based scheduling scheme, ihe scheduhrr i.ises the scores to prioritize terminals for lissignment to the available Mansmit antijnriis. The scores of the terminals are updated based on their assipments or non-assij;:DD:ii5nts. to transmit antennas. The active terminals in the list may be jiriorirized sji;h thai: the terminal with the lowest score is given the highest priority, and ihe terminal with nhe hij^est score is conversely given the lowest priority. Other jTiethods for ranking terminals may jdso be used. Tlic prioritization may also assign non-uniforrr weighting factors to the lerminal scores.
110811 For a. scheduling scheme in which terminals are selected and scheduled for data tr;msmi£;;:;ion based on their priority, it is possible for poor tenninal groupings to ' (>:cur c»ccasiona:.ly- A "poof terminal set is one that results, in similar channel response iriatrictJS Hit which cause similar and pr SNRs for all terminals on all transmit data ittreamii as given in the hypothesis matrix T. This then results in low overall throughput Tor ca:h tenrdnel in the set. When this happen, the priorities of the terminals may not diange subs:inljally over several frames. In this way, the scheduler may be stuck with this particular terminal set until the pricMities change sufficiently to cause a change in membership -n the set.
II082] To avoid the above-described "clustering'" ctltct, the scheduler can be
devised to :oxot the scheduler to select terminal sets that result in "good" hypothesis matrices (i.(^,
[1084] More complex scheduling schemes may also ])e devised that may be able to achieve throLi^ihput closer to optimum. These schemes may be required to evaluate a large)' number c-f h:/potheses and antenna assignments in oider to determine the best set of terriinals ;iui(i the best antenna assignments. Other sclieduling schemes may also be designed to t&e advantage of the statistical distribution of the data rates achieved by each tiirminfiJ. This information may be useful in reducing the number of hypotheses to fce evduated.. fn addition, for some applications, :t may be possible to learn which terminal groupings (i.e., hjpotheses) work well by analyzing performance over time. ITiis iifonriation may then be stored, updated, and used by the scheduler in future schediling in:ervals.
[1085] ';ite techniques described above may be used schedule terminals for data tiransrcission using the MIMO mc-de, N-SIMO mode, and mixed-mode. Other considsrations may also t>e applicable for each of these operating modes, as described beloM'.
MIMO Mode [1086] Iti tine MIMO mode, (up to) NT independent data streams may be simuUimeouiiily transmitted by the base station from Nr transmic antennas and targeted to a single M[I»
> Nj. Tae :entiinal may use spatial equalization (for a ron-dispersive MIMO channel ^^'ith a flat fietjuency chsmnel response) or space-time equalization (for a dispersive AHMC char nel with a frequency defendent channel response) to process and separate :he N'r trans:Titted data streams. The SNR of each post-proc:essed data stream (i.e., after equalization) may be estimated and sent back to the base station as CSI, which then uses "he intorma::ori to select the proper c
Ijcinsmi: antenna such thai: the target terminaJ i;; able to detect each transmitted data Tjeam at the deifired level of performance.
11087] If :iU data streams are transmitted to one temiinal as is the case in the MIMO
iiuode, tien tlie successive cancellation receiver processing technique may be used at the
(M-niinal to pnxiess NR received signals to recover NT Jransniitted data streams. This
kH^nique suci:
:!leratiCTis) to re::over the signals transmitted from the terminals, with one transmitted
i-l^nal leing nscovei-ed for each iteration. For each iteration, the technique performs
linear cr non-lirtear processing (i.e., spatial or space-time equalization) on the NR
::(iu:eive(l signals to recovei one of the transmitted signals, and cancels the interference
due to the nxcvered signal from the received signals to derive "modified" signals
having the ini:eri'e:rence component removed. i
l!bl)88] TIic: iciodified signals are rhtn processed b)' the next iteration to recover aiothei: trans:Tjii:ted signal. By removing the interference due to each recovered signal :iom ttic rccei\'ecl sij^als, the SNR improves for the transmitted signals included in the ir.(xiifiei sigruils but not yet recovered. The improved SNR results in improved ]»;rforr^ance i\n the terminal as well as the system, bi fact, under certain operating
[1089] In an emlwdiment, each MMO terminal in the system estimates and sends tack NT post-]:'r«>:essed SNR values for the NT transmit antennas. The SNRs from the active :enmni]Js may be evaluated b> die scheduler to c^iemaine which terminal to trtnsmit to aiui when, and the proper coding and modulalic^n scheme to use on a per triasmit anter.na basis for each selected terminal.
[1090] M!IM(I> terminals may be sel^ected for data trarsrnission based on a particular pstifonriiince maliic formulated to achieve the desired system goals. The performance
iietric: may b;5 based on one or more functions and any nunbcr of parametiars. Various functic-ns rany be used to formulate tlie performance metiic, such as the function of the ichievible throughput for the MIMO terminals, which is sliown above in equations (3)
lad (^.N -SIMP Mode ;;1091] If:i the N-SMO mode, (up to) NT independent data streams may be jjmuluineously transniitted by the t)ase station from tlie NT transmit antennas and iirget^d to (ii];> to) Nx different SIMO terminals. TC' maximize performance, the scheduler may consider a large numtcr of possible tennind sets for data transmission.
. Ihe scheduler *:i::ien detennines the best set of Nx tcrminalu to transmit simultaneously on a gpven crimael (i,e., time slot, co^ie channel, frequency sub-channel, and so on). In I multiple access conununication system, there are generally constraints on satisfying osrtain requ:.riij:rients on a per terminaJ basis, such as maximum latency or average data --ate. (n this c^ise, the scheduler can IK; designed to select the best set of terminals :iubjcci to \hcr^ constraints.
.:10921 I.i one implementation for the N-SIMO mode, the terminals use linear spatial .iquali5;ation lo process the receive signals, and the post-processed SNR corresponding .0 cacii transmit antenna is provided to the base station. The scheduler tiien uses the Information l^ select the terminals for data transmission and to assign the transnoit ititenii as to ihe 5^Iected terminals.
IM0931 I'l anotlier implementation for the N-SIMO mode, the terminals use ijucceiJiiive c-ancellation receiver processing to process tlie receive signal to achieved
. highe]: post-pr(;M:essed S^^Rs. With successive canceUation receiver processing, the ]>3st-piocessei SNRs for the transmitted data streams depend on the order in which the data streams m^t detected (i.e., demodulated and decode^d). In some cases, a particular ;>IMO' termir.i.:il may not bs able to cajicel the interference fiom a given transmitted df^t?^ stream intended for another terminal, since the coding emd modulation scheme used for Ihis data stroMr:i w£5 selected based on the other terminal's post-processed SNR. For nxamp e. th(?. ti:£insrnitted data stream may be targeted for terminal Ux and coded and modulated foi: proper detection at a (e.g., 10 dB) post-processed SNR achievable at the target termiriiil -Ux, but another terminsJ Uy may receive the same transmitted data stream ut a wcrse posit-processed SNR and is thus not able to properly detect the data stream. If
]ie data sticmi intendecl for another terminal caimcjt be detected error free, then uncelJation of the interference due to this data stream is not possible. Successive oancellation rsLjeivcr processing is viable when the post-processed SNR corresponding .0 a transmilKid datii stream permits rtiliable detection.
10941 I-- tJider for the scheduler to take advantage of the improvenient in post-5roces!;ed Sl
[1096] As note-d above, the transmit antennas may ht assigned to the selected letmirtiils baste on various schemes. In one antenna assignment scheme, the transmitted antennas are Jissigned to achieve high system performance and based on the priority of ihe teiTninalr,.
110971 Table 3 shows an example of the post-processed SNRs derived by each lermiriid in w hypothesis being considered. For terminal 1, the best SNR is achieved ^vhen detecting the data stream transmitted from transmit aritenna 3, as indicated by the iiiaded box m row 3, column 4 of the table. The bist transmit antennas for other termirjils in the liypothesis are also indicated by the shading in the boxes.
[1098j J: tiich tenninal identifies a different txanstnit Jinteima from which the best I)OSt-process
[109i>; ].f L'nore than one teraunal prefers the same transmit antenna, tlien the scheduler cm determine the antenna assignments bafjed on various criteria (e.g., faimc!ss, pe:rfoimance metric, and others). For example, Table 3 indicates that the best post-i3focesi>£:d SN!Rs for terminals 3 and 4 occur for the; data stream transmitted from the same tiaasjiiit antenna 1. If the objective is to nciaicimize throughput, then the scheduler nay as&ign tr.3nsmit antenna 1 to terminal 3 and transmit antenna 2 to terminal 4. tlowever, if antennas arc assigned to achieve fairness, then transmit antenna 1 may be assigned to terminal 4 if tenninal 4 has higher priority than terminal 3.
!Mixed-Modc [llOO] ".rhe techniques described above can be generalized to handle mixed SIMO and MDMO t<:3iaimds. for example if four transmit ajuenias are available at the base station theri rour independent data streams may be transmitted to a single mimo ti two terminals simo termijiiils one terminal plus t>vo IrA SIMO terrainals. or any other combination of terminals designated to niceivti a total of four da:a streams. The scheduler can be designed to select the best :ombinatior. of terminals based on tlie post-processed SMlis for various hypothesized sets of terminals, v/here each hjpothesized set may include a mixture of both MIMO indStvlO terminals.
;:1101] ^Vb;mever mixed-mode triiffic is supported, the use of successive .;.3nce]]ation ixceiver processing by the (e.g., MIMO) i:erminals places additional oonsti^ints Cf;:. the scheduler due to the dependencies introduced. These constraints may
result in metre hypothesized sets being evaluated, since in addition to considering :iffererit sct> of terminals tlie scheduler must also consider demodulation of the data ;tre;im.^ in viijious ijrders by each terminal. The asidgnment of the transmit antennas :nd the sele<:t of the coding and modulation schemes would then take into account :iese dipendeiici.es in order to achieve improved perfomiaiice.>
Transmit Antennas [1102] Th;; :5et of transmit antennas at a base station may be a physically distinct set A "apertures'*, each of wkich may be used to directly tr{ui;?mit a respective data stream, [uich aperture nay be formed by a collection of one or more antenna elements that are ;:istributed iri i3i:»acc (e.g., physically located at a single site or distributed over multiple 5i::es). Altenistively, the antenna ape;rtures may be preceded by one or more (fixed) :
i;U04] Ti<: use of beam-forming matrices affords additional flexibility in i tenninals and may further provide improved performance. as examples the r>Ilow.big sitiiiitions may he well suited for beam-forming transformations;
• Zorrelaton in the IvDMO channel is high so that the best performance may be achie'.'t:cl with a sm;il] number of data streams. However, transmitting with only a sut-iict of the available transniit antennas (and using only their associated xansrmt iunplifiers) results in a smaller total transmit power. A transformation :-nay ht sc;lected to use most or all of the transmit ajitennas (and their amplifiers)
for tn; (iata streams to be sent. In this case, higher traiismit power is achieved for tlu: transmitted data streams.
PhysK aliy dispersfid terminals may be isolated somewhat by their lixations. In this ciiSEi, the terminals may bt2 served by a standard FFT-type transformation of hori2;C'ni;iilIy space
l^erformance
111051 Tine techniquai described herein may be viewed as a particular form of itpatial division multiple access (SDMA) wherein each transmit antenna in the base station's anlcma aixay is used to transmit a different data stream using channel state information (e.g., SNRs or some other sufficient pfirsineter which determines the . i;upporlable data rate) derived by the terminals in the cc»verage area. High performance i is actiicved on the basiji of the CSI, which is useii in scheduling terminals and ])roceKi;ingdai:a.
|1106;i n^e i:echnique5. described herein can provide; improved system performance (e.g., h^ei throughput). Simulations have been perloiired to quantify the possible iystem throi:;jl:put with some of these techniques. In tlie simulations, the channel iesponr»e matrices Hi coupling the array of transmit antennas and the receive antennas of the .v-th toinilnal i& assamed to be composed of equul-Viiriance, zero-mean complex (jaussiin raidom variables. The simulations were performed for the MIMO and N-!>IMO modes.
11107] Li die MIMO node, four JvflMO terminals (each with four receive antennas) me coisideied for each realization (e.g., each transmission interval) and the best temainiil is i^eltxted and 5;chcduled fi)r data transmission. The scheduled terminal is iransni.tted four independent data streams and uses successive cancellation receiver (Ut)cesjing (withi MMSE equalization) to process the nsceived signals and recover the transtrctted data streams. The average tliroughput for the scheduled MIMO terminals is tecorded.
[1108] In tlie N-SIMO mode, four SIMO terminals, each with four receive antennas, iie coDsidered ]'ar each realization. ITie post-processed SlSRs for each SIMO terminal are dettjnnined using MM^SE linear spatial equalization (witliout successive cancellation receiver pro(:esii;ing). The transmit antennas are assigned to the selected terminals based
en the max-max criterion. The four scheduled leradnals are transmitted four independent iJaia streams and each terminal uses MNISE equaUzation to process the receive signal and recover its data stream. The throughputs for each scheduled SIMO tisrminal are s:epara*:ely recorded, and the average throughput for all scheduled terminals is also recorded.
[1109] F[G, 9 shows the average tliroughput for a J41M0 communication system v/ith four tnifismit antennjis (i.e., Nj == 4) and four receive aiuennas per terminal (i.e., NR =: 4) for the: MEMO and N-SMO modes. The simulatt^d throughput associated with sach (Operating mode is provided as a function of the average post-processed SNR, The avera.gs thr:>,ij;jiput for the MIMO mode is shov/n a;i plot 910, and the average tliroujjiput for the N-SIMO mode is shown as plot 912.
[1110] /^!> shovm in HG. 9, the simulated tiirougliput associated with the N-SMO mode jsing thi!; max-max criterion ajitenna assigmnent shows better performance than that achieved far the MIMD mode. In tlie MIMO mode, the MMO terminals benefit by jsing succeaiivc cancellation receiver processing to achieve higher post-processed SNRs. In tLe SD40 m«xle. the scheduling schemes are able to exploit multi-user selection divursily to acliieve improved performance (i.e., higher throughput) even iiough cacti SIMO terminal uses linear spatial equaiizarion, Jxi fact, the multi-user Jiveniity pro\'ided in the N-SIMO mode results in an average downlink throughput that exceeds the throughput achieved by dividing a transmission interval into four equal-iuraticn sut-slots and assigning each MIMO terminal to a respective sub-slot. ;illl] Trie scheduling schemes used in the simulations for both operating modes A'cre not desigxicd to provide proportionate fairness and some terminals will observe liigher average throughput than others. When a fairness criterion is imposed, the diffens ices in throughput for the two operating modes may diminish. Nevertheless, the .ibility to accommodate both MMO and N-SIMO terminajis provides added flexibility 10 the inrovisioriing of wireless data services.
|'][112] P'or simplicity, various aspects and embodiments of the invention have been dcscrited for a communication system in which (1) the number of receive antennas is i:^}ual to the numb^JT of transmit antennas (i.e., NR = Nx), and (2) one data stream is transmitted Irorn each antenna at the base station, in this case, the number of Iransm.ssion i:]:iiinnels is i^qual to the number of available spatial subchannels of the IvQMO cha-mel For a MEMO system that utilize C>FDM, multiple frequency
subchiinneli; nay be associated with each spatial sul>:hannel, and these frequency subchiinnels ixiiy be assigned to term nsds based on the techniques described above. For £. dispersive clumnel, a matrix H would represent a thjiieHlimensional cube of channel respor-se es.:iiT)£.tes for each terminal.
[1113] ]^'ic:h scheduled terminal :i\Ziy also be equipped with more receive antennas than tlie total ::iuml)er of data streams. Moreover, multiple terminals may share a give transjiiit anicniia, and the sharing may be achieved via time division multiplexing (e.g., assigning dfferent fractions of a transmssion interval to different tenninals), frequency division mL:':tipIexing (e.g., assigning different frequency subchannels to different terminals), codm division multiplexing (e.g., assigning different orthogonal codes to different terminals), some other multiplexing schemes;, or any combinations of these scheroes.
[1114' The scheduling schemes described herein select terminals and assign antenrasfor data tj-ansmission based on channel state infojmation (e.g., post-processed SlNRs]. The post-processed SNRs :~or the terminals are dependent on the particular transirit power level used for the data streams transmittiic from the base station. For simplicity, th^ same transmit power level is assumed for all data streams (i.e., no power control of the tranvsmit power). However, by controlling the transmit power for each antenca, th^ achievable SNRs may be adjusted. For exan^le, by decreasing the ti.^ansirit pow^ for a particular transmit antenna via power control, the SNR associated . v/ith a data slri^am transmitted from this antenna is reduced, the interference caused bv this ciita sti'e:an:i on other data streams would also be reduced, and other data streams may te able to achieve better SNRs Thus, power control may also be used in conjurction wjth the scheduling schemes described herein, and this is within the scope of the :nvendiD::i.
[1115] The- scheduling of terminals based on priority is .also described in U.S. Patent ^tpplication Serial No. 09/675,706, entitled "METHOD AND APPARATUS FOR DETERMINING AVAILABLE TRANSMIT POWJR IN A WIRELESS ::OMfdUNJ;CATION SYSTEM." filed September 29, 2000. Scheduling of data xansraission fcr the downlink is also described in U.S. Patent Application Serial No. 08/798,951, ectiUe^l "METHOD AND APPARATUS FOR FORWARD LINK RATE ,5CHEDULlh;'G;' filed St^tember 17, 1999. These applications are assigned to the itjsigntte of the present invention and incorporated herein by reference.
;ill6] Th:; scheduling schemes described herein incciporate a number of features 3rid provide numerous adviintages. Sonne of these features aid advantages are described
:t:low. 1117] F:T:jt., the schedujing schemes support various, operating modes, including
:iix-m(»de whi^jeby any combination of SMO imd MIMO terminals may be scheduled
:oY dala transiniiision on the downlinl^c. Each SIMO or MIMO terminal is associated
.^ith 2Ji SNl'rl: \'ectc'r (i.e., one row in equation (2)), The scheduling schemes can
jvaluatJ any number of possible combinations of termintils for data transmission.
;U18] S-K:cfid, the sclieduling scliemes pro^/ide a ischedule for each transmission
; itcrvd that inchides a set of (optimal or near optimal) "mutually compatible" terminals '
:£jsed on their spatial signatures. Mutual compatibility may be taken to mean
■:C'exist
ixnstrants regaining tenrdnals data rate requirements, transmit power, link margin,
i:apabil:ty betv/een SIMO and MIMO terminals, and possibly other factors.
111191 Third, the scheduling scheraes support variable data rate adaptation based on
:ii3 posl-procesiUKi SNRs iichieved at the terminals. Each scheduled terminal may be
informed when to expect data trtmsmiss-ion, the assigned ti'ansrait antenna(s), and the
iata rat2(s) Uy for the data transmission (e.g., on a per transmit antenna basis).
! 1120] Fourtli, the scheduling schemes can be designed to consider sets of terminals
!hat ha;e sii^iilar link margins. Termnals may be grouped according to their link
margin properij;s. The scheduler may then consider cc>mb:inations of tenninals in the
;ime "link mtigin" group when searching for mutually compatible spatial signatures.
'[Tiis gujuping according to link margin may improve the overall spectral efficiency of
I he scheduling schemes compared to that achieved by ignoring the link margins.
Moreover, by scheduling temiinals v/ith similar link mjirjpns to transmit, downlink
pcwer control rn.ay be mere easily exercised (e.g., on the entire set of terminals) to
:iiriprov(; ove:"sIl spectral rcuse. This may be viewed as a combination of a downlink
adaptive reus.e scheduling in combinadon with SOMA for SIMO/MD/IO. Scheduling
based oi link margins is described in further detail in U.S. Patent Application Serial No,
i)9/539,l57, ailitled "METHOD s^J^TD APPARAITJS FOR CONTROLLING
'TRANSMISSIONS OF A COMMUtnCATIONS SYSTIM," filed March 30, 2000,
i::id U.S. Paien: Application Serial No. [Attorney Docket No. PA010071], entitled
•^VDETl^OD AHl) APPAR.\TUS FOR CONTROLLING UI>LINK TRANSMISSIONS
OF A WKitLllSS COMMUNICATION SYSTEM/' filed May 3, 2001. both assigned to th
MIMO C oinmunication System 111211 KICr. 5 is a block diagrsira of base station KM- and terminals 106 within IvUMX) connnunication system 100, At base station 104, a data source 512 provides (lata (i.e., iifcrmation bi":s) to a trartsniit (TX) data processor 514. For each transmit anterjia, TX data processor 514 (1) encodes the dau in accordance with a particular codini; schinnt:^ (2) interleaves (i.e., recorders) the encoded data based on a particular intericavinj; ,:>c:heme, and (3) maps ttie interleaved bits into modulation symbols for one or more tranurnission chainnels selected for data tran£;rni.3slon. The encoding increases the nviabilit}' of the data transmission. The interleaving provides time diversity for the coded bits» permits the data to be transmitted based on au average SNR for the transmit fintenna, combats fading, and further removes correlation between coded bits used to ibrin each mcdulation symbol. The interleaving ma> further provide frequency diversity if die coded bits are transmitted over multiple frequency subchannels. In an iispeci, the coding and symbol mapping may be perforrricd based on control signals provicbd bj a scheduler 534.
[1122] Tlie encoding, interleaving, and signal mapping may be achieved based on various schemes. Some such schemes are described in the aforementioned U.S. Patent Application Serial No. [Attorney Docket No. PA010210]; U.S. Patent Application Serial No. 09/326,481, entitled "METHOD AND APPARATUS FOR UTILIZEs^G CHANNEL. STA'IE E^ORMATtON IN A WIREIJESS COMMUNICATION SYSTEM," filed March 23, 2001; md U.S. Patent Apphcation Serial No. 09/776,075, entithd "C:QDING SCHEME FOR A WIRED2SS COMMUNICATION," filed I'ebruary 1, 2001, all assigned to the assignee of tlie present application and incorporated herein by rcference.
[1123.' A. TX MIMO processor 520 receives and demultiplexes the modulation symbds from TX ilata processor 514 and provides a stream of modulation symbols for each transniisjiion channel (e.g., each transmit antenria}, one modulation symbol per liime slot. TX IvUMO processor 520 may further prscc>n*lition the modulation symbols for each sehi::ted transmission channel if full CSI (e.g., tlie channel response matrix H) is av£tilable. MIMO and i\ilI-CSI processing is described in further detail in U.S. Patent
AppIi<:atior iseriid no. entitled efficiency fflgh l comi system emptying multi-carrier modulatic filed march assigned to the assignee of present application aa incorporated herein by reference.>
I112^J] "f GFDM is not employed, TX MEMO processor 520 provides a stream of modulation i;>tcibols for each antenna used for data Ixansmission. And if OFDM is employed, TX MBVIO processor 520 provides a strejun of modulation symbol vectors • for ea:h antenna used for data transmission. And if fuIl-CSI processing is performed, "rX MIMO processor 520 provides a stream of preconditioned modulation symbols or precc»nditio::M;;d modulation symbol vxtors for each antenna used for data transmission. Each iitrcani is then received and modulated by a respective modulator (MOD) 522 and transmitted via an associated antenna 524. [112f}] A: each scheduled terminal 106, a number of receive antennas 552 receive
' the tninsmitleJ signals, and each leceive antenna provides a received signal to a respective deriiodulator (DEMOD) 554. Each demodulator (or front-end unit) 554 perforiis processing complenentary to that pcrfomed at modulator 522. The modulation symbols from all demodulators 554 are tlien provided to a receive (RX) filIMC>/data processor 556 and processed to recover one or more data streams transmitted for the terminal. RX MIMO/data processc^r 556 performs processing compl3men:ary to that performed by TX data processor 514 and TX MIMO processor 520 and provides decoded data to a data sink 560. Ttie p::ocessing by terminal 106 is clescril)ed in farther detail in the aforementioned U.S. Parent Application Serial Nos. [Attoiney Docket No. PI)010210] and 09/776,075.
[112(i] iVi: each active terminal 1C6, RX MIMO/data processor 556 further estimates tie Liik conditions and provides CSI (e.g., post-pK)ces3ed SNRs or channel gain estimates). A IX data pnxessor 562 then receives and processes the CSI, and provides processed data indicative of the CSI r.o one or more modulators 554. ModuIator(s) 554 farther concliion the processed data znd transmit the CSI back to base station 104 via a
:. revenjtJ cha:ij:iel. The CSI may be reported by the terminal using various signaling tijchniques ijs.g;., in full, differentially, or a combination thereof), as described in the aforementioned U.S. Patent Application Serial No. 09/826,481. [1127] M biise station 104, the transmitted feedback si;^al is received by antennas 524, d3modul.a:D3d by demodulators 522, and provided to a RX data/MIMO processor
'Si2. RX diiXa'MIMO processor 532 perfornis processing complementary to that ])iirfomed hy 'IX ciata p'ocessor 562 and recovers thie rcjported CSI, which is then ])::ovid(Ml to scheduler 534.
11128] Sclicdalex 534 uses the rejorted CSI to perfonn a number of functions such iLi (1) selecting ths set of best termnals for data transmission, (2) assigning the available traiiiimit antennas to the selected teiminals, and (3) determining the coding and jnodul^.tion jurheme to be used for eac:h assigned transmit ajitenna. Scheduler 534 may {.chedu.e terrciinals to achieve high throughput or based on some other ^performance (niteria or nieuijs, as described above. In FIG. 5, scheduler 534 is shown as being iraplenientec. wilhin base station 104. In other implementation, scheduler 534 may be iraplementec. within some other elenient of communication system 100 (e,g., a base station controlkr that couples to and interacts with a number of base stations). 11129J . FIG 6 is a block diagram of an embodiment of a base station 104x capable (rf processinj^ d;3ta for transmission to the terminals based, c^n CSI available to the base station (e.g.. ;3s reported by the terminals). Base stadon 104x is one embodiment of the transmtter port:ion of base station 104 in FIG. 5. Base stadon i04x includes (1) a TX ckta prxessor 514x that mceives and processes infomuitior bits to provide modulation sj^Dbols and (2) a TX IvCMO processor 520x that demultiplexes the modulation 5 jonbols for ':l:e N7 transmit antennas.
[1130] In tie specific embodiment shown in FIG. 6. TX data processor 5l4x includes a dtjraultip'exer 608 coupled to a number of channel data processors 610. one [Mocessor for each of the Nc transmission channels. Demultiplexer 608 receives and cbmultplexcs tiie aggregai;e information bits into a nun:iber of (up to Nc) data streams, one data streaiii for each cf the transmission channels to be used for data transmission, liach cLita streiut. is provided to a respective channel data prccessor 610. [1131j In ti'ie embodiment shown in FIG. 6, each ciiaimel data processor 610 includss an encoder 612, a channel imerleaver 614, and a s^nnbol mapping element 616. [incoder 612 r-eceives and encodes the information bits in ^iie received data stream in accordance v^-itlr a pardcuLar coding scheme to provide coded bits. Channel interleaver !']4 intorleaveis tie coded bits based on a particular interleaving scheme to provide time diversity. And symbol mapping element 616 maps the interleaved bits into modulation 3>mbol5 for the transmission channel used for transmitting the data stream.
[1132] F'lJot data (e.g.^ data of known pattern) may also be encoded and multiplexed g^'ith tte processed information bits. Tl-ie processed pilot dita may be transmitted (e,g., in a tine division multiplexed (TDM) manner) in iill or a subset of the transmission :lianncls ustxi to transmit the infonnation bits. The pilot data may be used at the xrminiils to pi^rform channel estimation.
;;1133] /t.s. l:;]■^o^^'n in FIG. 6, the data encoding, interfeiving, and modulation (or a combination thereof) may be adjusted based on the aviulable CSI (e.g., as reported by iie tenninai:0- iji one coding and moilulation scheme, adaptive encoding is achieved by ising u fixed base code (e.g.. a rate 1/3 Turbo code) and adjusting the puncturing to iichie\'(5 the dcsij'ed code rate, as supp:)rted by the SNR of the transmission channel used ':o transmit the data. For this scheme, the puncturing may be performed after the •:hannel inttn-leaving. In another coding and modulation scheme, different coding :;chemes may bi: us(5d based on the re]X)ited CSI. -For example, each of the data streams nay tKs codtrd ^^dth an independent c^de. With this scheme, a successive cancellation :rx:eivcr proct-Sijing scheme may be used at the terminals to detect and decode the data i.lreain;; to derive a more reliable estimate of the transnriitt&l data streams, as described : n furtl er de .ail l^elow.
[1134] SymlK)! mapping element 616 can be designed to group sets of interleaved l>its tC' form non-binary s)Tnbols, and to map each non-binary symbol into a point in a signal :onsttjll£.tion corresponding to a particular modulation scheme (e.g., QPSK, M-li^SK, J.1-QAM, or some other scheme) selected for tlie ti'ansmission channel. Each mapped signal point corresponds to a modulation symbol The number of information b:.ts th.it m£.y be transmitted for each modulation symbc»l for a particular level of performance (e.g., one percent packet enor rate (PER)) is dependent on the SNR of the transmission chfjinel. Thus, the coding and modulation jscheme for each transmission channel may t'C selected based on the available CSL The channel interleaving may also 1x5 adjusted b;isai on the available CSI*
[1135; The modulation symbols from TX data processor 514x are provided to TX t/OMO proce;3sor 520x, wliich is one embodiment of TX MEMO processor 520 in FIG. IK 'ft^.thin T>;: MMO inrocessor 520x, a demultiplcster 622 receives (up to) Nc tnodulation symbol streams from Nc channel data processors 610 and demultiplexes the received mo
symbC'l streaii;. as provided to a respeciive modulator 522. Each modulator 522 converts the modulation symbols into an analog signal, and further Simplifies, filters, quadrature modulates, and upconverts tlie signal to generate a nriodulated signal suitable for transmission civer the wire'ess link.
[1136] A transmitter design tliat implemenbi OFDM is described in the £ilbremintionsd U.S. Patent Application Serial Nos, [Atl:oimey Docket No. PA010210L 09/826 481. ;wy776.075, and 09/532,492.
[1137]! FKJ, 7 is a block diagrara. of an embodiment cf terminal 106x capable of inpleir.enting v axious aspetcts and embodiments of the invention. Terminal 106x is one stmbocliment of the receive portion of terminals 106£L thj-ough 106n in FIG. 5 and impleir.ents the successive cancellation receiver processing technique to receive and recovei the li*a:nsniitted signals. The transmitted signals from (up to) NT transmit ^iiitenriis are received by each of NR antennas 552a through 552r and routed to a tespect.ve demodulator (DEMOD) 554 (which is also referred to as a front-end [processor). Each dismodulator 554 conditions (e.g., fil:ei"s and amplifies) a respective teceivei signal, downconverts the conditioned signal to an intermediate fiequency or l>aseband, and digitizes the downconvcrtcd signal to provide samples. Each deOTodulator 554 may further demodulate the samples w itli a received pilot to generate a stream of received modulation symbols, which is provided to a RX MIMO/data [iroccssor555A.
[1138] In ±t embodiment shown in FIG. 7, RX MDvlO/data processor 556x (which is one ombo-iment of RX MIMO/data processor 556 in FIG. 5) includes a number of successive (i.e., cascaded) receiver processing stages 710, one stage for each of the tninsiritted ikm stre^im to be recovered by terminal I06x. In one transmit processing =cheme, one data stream is transmitted on each transmission channel assigned to t5rmin£l I06x., ;ind each data stream, is independently pn>cessed (e;g., with its own coding and rrjoduladon sc:heme) and transmitted from a r^pective transmit antenna. Por thii. transioit processing scheme, the number of data s'xeams is equal to the number :A' assi.jned irarismission channels, which is also equal i:o the number of transmit sjitennis assigned for data transmission to terminal 106x (which may be a subset of the SA'ailab.e trai:i5;mit antennas). For claiity, RX MIMO/d;ita i)rocessor 556x is described for thiij trans:iiit processing scheme.
11139] ::;.iK:h receiver processing stage 710 (except for Lie last stage 710n) includes a chann-Jl MItvIO'data prodjssor 720 coupled to an interference canceller 730, and the last 5:tage 710n includes onJ> channel MEMO/data procesi>or 720n. For the first receiver processing 3tai;e 710a, channel MIN![0/data prcicessor 720a receives and processes the NR ni3dulaLi'3::i synbol streams from demodulators 554a through 554r to provide a decod-;d dalS; i-lxeam for the first transmission channel (or the first transmitted signal). And for each cf the second through last stages 710b thi:oug;h 710n, channel MIMO/data processor 720 for that stage receives imd processes tlie NR modified symbol streams from 'he irly that stage. Each chaimel ■ l^CMO/data prtxessor 720 further provides CSI (e.g., t!ie SNR) lor the associated transmission channel.
[114(1] ]^or the first receiver pnxessing stage 710£i, interference canceller 730a receiV'js the S^, m-Ddulation sjinbol streams from all NR demodulators 554. And for each cf the sec c»nd through second-to-last stages, interference canceller 730 receives the NR m
[1141] In FIG- 7, a controller 740 is shown couplt^d to RX MIMO/data processor 556x iind may be used to direct various steps in tlie succ:essive cancellation receiver processing jH2rl:ormed by processor 556x.
[1142] FIG'. 7 shows a receiver structure that may 1>5 used in a straightforward manner whe;:: i^ach data stream is tr£insmitted over a resptJCtive transmit antenna (i.e., one data strsimi coxresponding to each transmitted signail). In this case, each receiver procciSjing stage 710 may be operated to recover one of the transmitted signals and provid'i the cteoded data stream comjsponding to the recovered transmitted signal. For
some other ixmiismi-: processing schemes, a data stream may be transmitted over multiple transirit anU;n:nas, frequency subchannels, and^or time intervals to provide spatial, frequeacy, and time diversity, respectively. For these schemes, the receiver processing i:aitial.ly deri ves; a r
[1143] ' ¥IQ. 8A is a block diajxam of an cmbodroicnt of channel MIMO/data processor 7?-0)(., which is one embodijnent of channel MIMO/data processor 720 in FIG. 7. In this tixabodimcnt, channel MDVIO/data processor 720x includes a spatial/space-tme processor 810, a CSI processor 812, a selector 814, a demodulation element 818, a de-intfTleavsi* 818, and a decoder 820.
[1144] Spacal/spacc-time processor 810 performs linear spatial processing on the ^[R re:sived si^pials for a non-dispersive MMO channel (i.e., with flat fading) or space-dme proces;5iQg on the Na received signals for a dispersive MIMO channel (i.e., with frequeicy s<:lective fading the spatial processing may be achieved using linear proce33ing techniques such as a chamiel correlation matrix inversion technique minimum :raejm square error and others. these usi to null out tfie undesired signjils or maximize received snr of each la in presence noise interference from other signals. space-time ai mivlse e decision feedback equalizer maxirrum-likclihocd sequence estimator ccmi mmse nimsk-le ami dfe teclmiques are described furtlisr detail aforementioned lr.s. patent serial no. docket pa010210 mlse techuiqaes also further by s.l. ariyavistakul et al. paper entitled>ptimum Space-Time Processors with Dispersive Interference: Unified 4.nal>5is and Riquired Filter Span," EEiEE Trans, on Conmiunication, Vol. 7, No. 7, July 1999, aid :nco:rporated herein by reference.
11451 C;SI processor S12 determines the CSI for each of the transmission channels ised for da:a Iransmission. For eximiple, CSI processor 812 may estimate a noise
covariance niatdx based on the received pilot signals and then compute the SNR of the />th ti£nsinir>5ion channel used for thie data stream to be decoded. The SNR can be £;4;timatsd similar to conventional pilcU assisted single md multi-carrier systems, as is iDiowTi in tl)e 5irt. The SNR for ;ill of the transmission channels used for data transmission may comprise the CSI that is reported back to the base station for this transmission channel. CSI processoi: 812 further provides to selector 814 a control signal that id.e:n.tifies the particular diita stream to l?e recovered by this receiver pjocessing stage.
I].146;i Selector 814 receives a ntmber of symbol streams from spatial/space-time [jrocessor 810 md Tnboh are tien provided to a channel simulator 830. which processes the symbols
'.vith ths esti:r.al:tJd channel response to provide estimates, i , of the interference due the decoded daui sixeanx. The channel response estimate may be derived based on the pilot and/or data tr£.nsniitted by the base station and in accordance with the techniques described in io^ aforementioned U.S. Patent Application Serial No. [Attorney Docket
]so,R4.0102W>:i.
/. Jfc |]1149J 11)5 N^ elements in the interference vector i correspond to the component
of the rcceivi2<:l si at each of the nr receive anttjnnas due to symbol stream>
iransnrtted on the fc-th transmit antenna. Each element of the vector represents an
(5i>tim£iled cctriiponent due to the decoded data stream in the corresponding received
jrioduls.tion .sjTiibol stream. These components are interference to the remaining (not
yet detected) trjiu-istrdtted signals in the NR received modulation symbol streams (i.e., the
vector E ), iirid are subtracted (i.e., canceled) from the n^ceived signal vector r by a
summer 832 to provide a modified ve<:tor r having the components from decoded>
ilita stieam ieB:ioved. The modified vector r*^^ is provided as the input vector to the
next rcceiva' i:)rccessing stage, as shown in HG. 7.
13.150] Various aspec.s of the successive cancellation receiver processing are
(iescribed in ili^rlher detail in the afcrenaentioned U.S. Pai:ent Application Serial No.
I Attorney D(K:ket No. PA010210].
11151] Receiver designs that do not employ the successive cancellation receiver
processing te:h:nique may also be used to receive, process, and recover the transmitted
cL'ita ffb-eams. Some such receiver designs are described in the aforementioned U.S.
I^iitent Application Serial Nos. 09/776,075 and 09/826,481. and US. Patent Application
Serial No. i:'9;:532.492, entitled '^MGH EFHCIENCY. fflGH PERFORMANCE
COMIvrUNKlAnONS SYSTEMENfPLOYING N0:TI-C.\RRIER MODULATION,"
Itled March 30, 2000, assigned to the ;assignee of the present invention and incorporated
tterBin])yrefe;rence,
[11521 F"»-i" simplicity, various aspec':s and embodiments of the invention have been
cictscribsd whenitin tlie CSI comprises SNR. In general, ths CSI may comprise any type
:»f informaticm that is indicative of the characteristics of the communication link.
Various typc!5 o! information may be provided as CSI, some examples of which are
:lc;scril>2dbe'.ow.
11153] III one embodiment, the CSl coniprisas sig?iiil-to-noise-plus-interference lEtio (SNR): '^'hich is derived as the ratio of the signal power over the noise plus interference po\ver. The SNR is typically estimated
11155] In yet another embodiment, the CSL comprises signal power, interference j>ower, and r
11156] bi yet another embodiment, the CSI comprises signal-to-noise ratio plus a h:st of interference powers for each observable interference term. This infonnation may Ixi derived and provided for each transmission channel used for data transmission. 11157] In yei: another embodiment, tlie CSI comprises signal components in a matrix lonn M5.g,, t>T5
noise plus interference components in-matrix form (e.g., N^xNj^ complex entries).
~.7hQ ba>c statiiDn may then properly combine the signal components and the noise plus interference components for the appropriate transmit-receive antenna pairs to derive the c[uality for each transmission channel used for data trEJismission (e.g., the post-{>rocessed SNR for each transmitted data stream, as received at the terminals). [1158] In 'yt^t another embodiment, the CSI comprises a data rate indicator for each tninsnriit.dati stream. The quality of a transmission channel to be used for data tnmsffiission m;iy be determined initially (e,g., based on the SNR estimated for the • nmsnriission channel) and a data rate corresponding to the determined channel quality caay tlisn be identified (e.g., based on a look-up table). The identified data rate is indicadve of tine maximum data rate that may be transmitted on the transmission :hanne; for ric required level of performance. The diOa rate is then mapped to and represcited by a data rate indicator (DIRT), which can be efilcientiy coded. For example, : :f (up t^) sev£;n possible data rates ars supported by the base station for each transmit
ititenna, then i\ 3-bit value may be used to represent tlie DRI where, e.g., a zero may ndicats a davi rate of zero (i.e., don't use the transmit antenna) and I through 7 may be iscd to indicate: seven different data rates. In a typical implementation, the quality jieasuremerts (e.g., SNR estimates) are mapped directly to the DRI based on, e^g., a ook-u]3tabh.
;:11S9] In £iriother embodiment, the CSI comprises power control information for iach transmission channel The power control infonnation may include a single bit for jach tiansmiission chaimel to indicate a request for either more power or less power, or t may include multiple bits to indic:ate the raagnitudci of the change of power level :355que!iied Irii tiiis embodiment, the base station may mal«:e use of the power control ' : riformation fed back from, tlie terminzils to adjust the data piocessing and/or the transmit
1 >ower. ;il60I L:i yet ar other embodimcni, the CSI comprises an indication of the particular jfocejiiiing scheme to be used at the base station for each transmit data stream. In this «i:iibo2rfomance is achieved.
[11611 i"! yet another embodiment, the CSI compnses a diffcrcntiaJ indicator for a ]3articular rae^ascire of quality for a transmission channel Initially, the SNR or DRI or some 3ther quiditj' measurement for the transmission channel is determined and itiported as a :re ference measurement value. Thereafter, monitoring of the quality of the iransm.ssion channel continues, and the difference between Lhe last reported ineasinement and the current measurement is determined. The difference may then be qaantized to one or more bits, and the quantized difference is mapped to and represented by the differential indicator, which is then reported. The differential indicator mjiy indicate to increase or decrease the last reported measurement by a ]);irticular step size (or to maintain tlie last reported measurement). For example, the (Lffercntial indicator may indicate ttiat (1) the observed SNR for a particular Jransniission channel has increased or decreased by a particular step size, or (2) the data r£;te should be adjusted by a particuhir amount, or some other change. The reference iceasurement n:iay be transmitted peiiodically to ensure that errors in the differential irdicators and'or cnoneoujj reception of these indicators do not accumulate.
11162] Cvtiier fcrms of CSI may also be used and ^:e within the scope of the invent! Dn. Ji tjeneral, the CSI includes sufficient iofonnation in whatever fonn that . ir;ay bt. used i:o adjust the processing at the base station such thiat the desired level of perfornance is i3chieved for the transmitted data streams.
11163] The CSI may be derived based on the signals transmitted from the base jtation and rex:eived at the terminals. In an embodiment, the CSI is derived based on a pilot reference: included in the transmitted signals. Alternatively or additionally, the CSI inay be deri\ tid based on die data included in the transmitted signals. 11164] In yiit imothex embodiment, the CSI comprises one or more signals transmitted on the liplink from the terminals to the base station. In some systems, a (Uigree of co;:i:^;Iaiicn may exist between the uplink and downlink (e.g. time division (luplexiid (TOD) systems where the uplink and downlinlc share the same band in a time division multiplexed manner). In tht^se systems, the quEiIity of the downlink may be £ s-timatsd (tct \i requisite degree of accui'acy) based on the quality of the uplink, which may be estiinited based on signals (e.g., pilot signals) trimsmitted from the terminals. "jie pibi signiilii wculd then represent a means for which the base station could estimate the CSI as obser/ed at the terminals.
[1165] Th:; signal quality may te estimated at the terminals based on various t^hniques. Some of these: techniques aie described in the following patents, which are assigned to tiie assigiiee of the present application and in'Xrporated herein by reference:
• U.S. Pa:ent No. 5,799,005, entitled "SYSTEM AND METHOD FOR DETIERMINING RECEIVED PILOT POWER A^) PATH LOSS IN A CDM A COMMUNICATION SYSTEM," issued Aujjust 25,1998,
• U.S. Patsnt No. 5,903,554. entitled "METHOD AND APPARATUS FOR MEAS URING IJNK QUALITY IN A SPREAD SPECTRUM COMlvnjNICATION SYSTEM/^ issued May 11> 1999,
• U.S. Faient Nos. 5,056.109, and 5,265.119, bDtli entitled "METHOD AND APPAR-^.TUS FOFL C0NTR0:LLING TRANSMISSION POWER EN A CDMA ::ELIJ]LAR MOBE-E TELEI^HONE SYSTEM," respectively issued October E, 1991 iiind November 23,1993, and
• U.S. Patent No. 6,097,972, entitled "METHOD AND APPARATUS FOR
pRO'::ESSI^ra POWER COI^TROL SIGNAUS IN CDMA MOBILE
TEUHPHONE SYSTEM," issued August 1.2000.
[/.:ethc
[1166] V liri'ous types of information for CSI and vaiiouij CSI reporting mechanisms Eie al^o d£:;3cnbed in U.S. Patent Application Serial No. 08/963,386, entitled I^IETilOD AND' AFP.\RATUS FOR HIGH RATE PACKET DATA ]:ElAhf;)MISSION;' filed November 3, 1997, assigned to the assignee of the present applicadon, aiii] in *TIE/]3IA/IS-8S6 cdma2000 High Rate; Packet Data Air Interface Specifioation'% both of wh:ch are incorporated herein by reference. [1167] Thi3 CSI may be r^iported back to the base station using various CSI tnmsDriission ^.chemes- For exanple, the CSI may be sent in full, differentially, or a combination tiioreof. In one embodiment, CSI is reported i)eriodically, and differential ipdatcs are ^eat based on the prior transmitted CSI. In another embodimeni, the CSI is sent only when the?e is a change (e.g., if the change exceeds a particular threshold), v/hich ]nay I:>wer the effective rate of the feedback channel As an example, the SNRs raay be sent back (e.g., differentially) only when they change. For an OFDM system [vAth cr without MIMO). correlation in the frequency domain may be exploited to fi€jrmit reduction in the amount of CSI to be fed back. As an example for an OFDM system, if the SIMR corresponding to a particular spatigj .subchannel for NM frequency subchannels is the same, ttie SNR and the first and last frequency subchannels for which ■ :iis condition is true may be reported. Other compression and feedback channel error recovery teamiques to reduce the amount of data to be fed back for CSI raay also be :ied and axe v/itlidn the scope of the invention.
;il68] Ths 'siements of the base station and terminals may be implemented with one •:r moK: digi'.il signal processors (DSP), application spe:iric integrated circuits (ASIC), ;:rocessors, :n:iicroprocessors, controllers, microcontrollers, field programmable gate
iUTays yPGcS). programrriable logic devices^ other electronic units, or any combination ttiereof. SoTie of the functions and processing clesciibed herein may also be itnplenientecl with software executed on a processor.
11169] C:-;n:an Jispects of the invention may be implemented with a combination of j.oftwiue and hardware. For example, the processing to 5;chedule (i.e., select terminals imd assign tran!;:niit antennas) may be performed based on program codes executed on a processor (scheduler 534 in FIG. 5).
I1170J Headings are included herein for reference and to aid in the locating certain 5 ections. Tlu:se heading jare not intended to limit the scope of the concepts described therein under, and these concepts ma^y have applicabihty in other sections tliroughout the entire sp»:;dlicati.on.
11171J The previous description of tlie disclosed embodiments is provided to enable i\i\y person tikillec in tlie art to mzike or use the present invention. Various icodificaitions to these embodiments will be readily apparent to those skilled in the art, ciiid the generic jrirciples defined herein may be applied to other embodiments without :[eparti:ig fn»ra Ihc spirit or scope of the invention. Thus, the present invention is not intended to t»e. ]:i:iuted to thie embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
[1172] VrBAT B CL^UMED IS:
[■ ■.
WE CLAIM :
1. A method for scheduling downlink data transmission to a plurality of terminals in a wireless communication system, comprising: forming one or more sets of terminals for possible data transmission, wherein each set includes a combination of one or more terminals and corresponds to a hypothesis to be evaluated; assigning a plurality of transmit antennas to the one or more terminals in each set; evaluating performance of each hypothesis based in part on antenna assignments for the hypothesis; selecting one of the one or more evaluated hypotheses based on the performance; and scheduling data transmission to the one or more terminals in the selected hypothesis.
2. The method of claim 1, further comprising: forming a plurality of sub-hypotheses for each hypothesis, wherein each sub-hypothesis corresponds to specific assignments of the transmit antennas to the one or more terminals in the hypothesis, and wherein the performance of each sub-hypothesis is evaluated and one of the evaluated sub-hypotheses is selected based on the performance.
3. The method of claim 1, wherein the assigning includes identifying a transmit antenna and terminal pair with a best performance among all unassigned transmit antennas, assigning the transmit antenna in the pair to the terminal in the pair, and removing the assigned transmit antenna and terminal from consideration.
4. The method of claim 1, wherein each hypothesis is evaluated based in part on channel state information (CSI) for each terminal in the hypothesis, wherein the CSI is indicative of channel characteristics between the transmit antennas and the terminal.
5. The method of claim 4, wherein the CSI for each terminal comprise signal-to-noise-plus-interference ratio (SNR) estimates derived at the terminal based on signals transmitted from the transmit antennas.
The method of claim 5, wherein each set of one or more terminals to be evaluated is associated with a respective matrix of SNRs achieved by the one or more terminals in the set.
The method of claim 5, wherein one or more sets of antenna beams are evaluated by each terminal to be considered for scheduling to provide one or more vectors of SNRs, one vector for each set of antenna beams.
The method of claim 4, further comprising: determining a coding and modulation scheme for each transmit antenna based on the CSI associated with the transmit antenna.
The method of claim 1, wherein the one or more terminals in each set are selected from among a pool of terminals.
The method of claim 9, wherein the pool of terminals includes one or more SIMO terminals each designated to receive a single data stream.
The method of claim 10, wherein the selected hypothesis includes a plurality of SIMO terminals.
The method of claim 8, wherein the pool of terminals includes one or more MIMO terminals each designated to receive multiple data streams from multiple transmit antennas.
The method of claim 12, wherein the selected hypothesis includes a single MIMO terminal.
The method of claim 12, wherein each scheduled MIMO terminal performs successive cancellation receiver processing to recover data transmitted to the MIMO terminal.
The method of claim 1, wherein each set includes terminals having similar link margins.
The method of claim 1, wherein the evaluating includes computing a performance metric for each hypothesis.
The method of claim 16, wherein the performance metric is a function of throughput achievable by each terminal in the hypothesis.
The method of claim 16, wherein the hypothesis having the best performance metric is selected for scheduling.
The method of claim 1, further comprising: prioritizing terminals to be considered for scheduling.
The method of claim 19, wherein the plurality of transmit antennas are assigned to the one or more terminals in each set based on the priority of the terminals in the set.
The method of claim 20, wherein a highest priority terminal in the set is assigned a transmit antenna associated with a highest throughput, and a lowest priority terminal in the set is assigned a transmit antenna associated with a lowest throughput.
The method of claim 19, further comprising: limiting terminals to be considered for scheduling to a group of N highest priority terminals, where N is one or greater.
The method of claim 19, further comprising: maintaining one or more metrics for each terminal to be considered for scheduling, and wherein the priority of each terminal is determined based in part on the one or more metrics maintained for the terminal.
The method of claim 23, wherein one metric maintained for each terminal relates to an average throughput rate achieved by the terminal.
The method of claim 19, wherein the priority of each terminal is further determined based on one or more factors maintained for the terminal and associated with quality of service (QoS).
The method of claim 1, wherein the one or more terminals in the selected hypothesis are scheduled for data transmission over a channel that includes plurality of spatial subchannels.
The method of claim 1, wherein the one or more terminals in the selected hypothesis are scheduled for data transmission over a channel that includes plurality of frequency subchannels.
A method for scheduling data transmission to a plurality of terminals in a wireless communication system, comprising: forming one or more sets of terminals for possible data transmission, wherein each set includes a unique combination of one or more terminals and corresponds to a hypothesis to be evaluated; forming one or more sub-hypotheses for each hypothesis, wherein each sub-hypothesis corresponds to specific assignments of a plurality of transmit antennas to the one or more terminals in the hypothesis; evaluating performance of each sub-hypothesis; selecting one of a plurality of evaluated sub-hypotheses based on their performance; scheduling data transmission to the one or more terminals in the selected sub-hypothesis; and transmitting data to each scheduled terminal in the selected sub-hypothesis from one or more transmit antennas assigned to the terminal.
The method of claim 28, wherein the evaluating includes determining a throughput for the one or more terminals in the sub-hypothesis based on the specific antenna assignments, and wherein the sub-hypothesis with highest throughput is selected.
The method of claim 28, wherein one set of terminals is formed, and wherein the terminals in the set is selected based on priority of terminals desiring data transmission.
A multiple-input multiple-output (MIMO) communication system, comprising: a base station comprising a plurality of transmit antennas, a scheduler configured to receive channel state information (OSI) indicative of channel estimates for a plurality of terminals in the communication system, select a set of one or more terminals for data transmission on a downlink based at least in part on the received OSI, and assign the plurality of transmit antennas to the one or more selected terminals, a transmit data processor configured receive and process data for the one or more selected terminals based on the CSI to provide a plurality of data streams, and a plurality of modulators configured to process the plurality of data streams to provide a plurality of modulated signals suitable for transmission from the plurality of transmit antennas; and one or more terminals, each terminal comprising: a plurality of receive antennas, each receive antenna configured to receive the plurality of modulated signals transmitted from the base station, a plurality of front-end units, each front-end unit configured to process a signal from an associated received antenna to provide a respective received signal, a receive processor configured to process a plurality of received signals from the plurality of front-end units to provide one or more decoded data streams, and to further derive OSI for the plurality of modulated signals, and a transmit data processor configured to process the CSI for transmission back to the base station.
A base station in a multiple-input multiple-output (MIMO) communication system, comprising: a transmit data processor configured to receive and process data to provide a plurality of data streams for transmission to one or more terminals scheduled for data transmission, wherein the data is processed based on channel state information (OSI) indicative of channel estimates for the one or more scheduled terminals; a plurality of modulators configured to process the plurality of data streams to provide a plurality of modulated signals; a plurality of transmit antennas configured to receive and transmit the plurality of modulated signals to the one or more scheduled terminals; and a scheduler configured to receive OSI for a plurality of terminals in the communication
system, select a set of one or more terminals for data transmission based at least in part on the received CSI, and assign the plurality of transmit antennas to the one or more selected terminals.
The base station of claim 32, wherein the data stream for each transmit antenna is processed based on a coding and modulation scheme selected for the transmit antenna based on the CSI associated with the transmit antenna.
The base station of claim 32, further comprising: a plurality of demodulators configured to process a plurality of signals received via the plurality of transmit antennas to provide a plurality of received signals, and a receive data processor configured to further process the plurality of received signals to derive CSI for the plurality of terminals in the communication system.
A terminal in a multiple-input multiple-output (MIMO) communication system, comprising: a plurality of receive antennas, each receive antenna configured to receive a plurality of modulated signals transmitted from a base station; a plurality of front-end units, each front-end unit configured to process a signal from an associated received antenna to provide a respective received signal; a receive processor configured to process a plurality of received signals from the plurality of front-end units to provide one or more decoded data streams, and to further derive channel state information (CSI) for each decoded data stream; and a transmit data processor configured to process the CSI for transmission back to the base station, and wherein the terminal is one of one or more terminals included in a set scheduled to receive data transmission from the base station in a particular time interval, and wherein the set of one or more terminals scheduled to receive data transmission is selected from among one or more sets of terminals based at least in part on OSI received from the one or more terminals in each set.
The terminal of claim 35, wherein the terminal is scheduled to receive data transmission from one or more transmit antennas at the base station assigned to the terminal.
An apparatus for scheduling data transmission to a plurality of terminals in a wireless communication system, comprising: means for forming one or more sets of terminals for possible data transmission, wherein each set includes a combination of one or more terminals and corresponds to a hypothesis to be evaluated; means for assigning a plurality of transmit antennas to the one or more terminals in each set; means for evaluating performance of each hypothesis based in part on antenna assignments for the hypothesis; means for selecting one of the one or more evaluated hypotheses based on the performance; and means for scheduling data transmission to the one or more terminals in the selected hypothesis.
The apparatus of claim 37, further comprising: means for forming a plurality of sub-hypotheses for each hypothesis, wherein each sub-hypothesis corresponds to specific assignments of the transmit antennas to the one or more terminals in the hypothesis, and wherein the performance of each sub-hypothesis is evaluated and one of the evaluated sub-hypotheses is selected based on the performance.
The apparatus of claim 37, wherein means for assigning includes: means for identifying a transmit antenna and terminal pair with a best performance among all unassigned transmit antennas; means for assigning the transmit antenna in the pair to the terminal in the pair; and means for removing the assigned transmit antenna and terminal from consideration.
The apparatus of claim 37, wherein each hypothesis is evaluated based in part on channel state information (OSI) for each terminal in the hypothesis, wherein the OSI is indicative of channel characteristics between the transmit antennas and the terminal.
The apparatus of claim 40, wherein the OSI for each terminal comprise signal-to-noise-plus-interference ratio (SNR) estimates derived at the terminal based on signals transmitted from the transmit antennas.
The apparatus of claim 40, further comprising: means for determining a coding and modulation scheme for each transmit antenna based on the OSI associated with the transmit antenna.
The apparatus of claim 42, wherein means for evaluating includes: means for computing a performance metric for each hypothesis.
The apparatus of claim 43, wherein the performance metric is a function of throughput achievable by each terminal in the hypothesis.
The apparatus of claim 37, further comprising: means for prioritizing terminals to be considered for scheduling.
The apparatus of claim 45, wherein the plurality of transmit antennas are assigned to the one or more terminals in each set based on the priority of the terminals in the set.
The apparatus of claim 45, further comprising: means for maintaining one or more metrics for each terminal to be considered for scheduling, wherein the priority of each terminal is determined based in part on the one or more metrics maintained for the terminal
An apparatus for scheduling data transmission to a plurality of terminals in a wireless communication system, comprising: means for forming one or more sets of terminals for possible data transmission, wherein each set includes a unique combination of one or more terminals and corresponds to a hypothesis to be evaluated; means for forming one or more sub-hypotheses for each hypothesis, wherein each sub-hypothesis corresponds to specific assignments of a plurality of transmit antennas to the one or more terminals in the
hypothesis; means for evaluating performance of each sub-hypothesis; means for selecting one of a plurality of evaluated sub-hypotheses based on their performance; means for scheduling data transmission to the one or more terminals in the selected sub-hypothesis; and means for transmitting data to each scheduled terminal in the selected sub-hypothesis from one or more transmit antennas assigned to the terminal.
The apparatus of claim 48, wherein means for evaluating includes: means for determining a throughput for the one or more terminals in the sub-hypothesis based on the specific antenna assignments, and wherein the sub-hypothesis with highest throughput is selected.
|