CA2109789C - Diversity for direct-sequence spread spectrum systems - Google Patents

Abstract

An invention for providing diversity for direct sequence spread spectrum wireless communication systems is presented. The invention provides a transmitting technique for communicating a first signal comprising one or more segments to a receiver with use of a plurality of M antennas. The first signal includes one or more signal segments. The technique comprises forming M copies of the first signal; for a segment of a signal copy, weighting each of two or more sub-segments of the segment with a distinct signal, wherein a sequence of the distinct weighting signals for the segment is distinct from sequences of signalsweighting the same segment of one or more other signal copies; and for each of Mweighted signal copies, transmitting a signal to the receiver using a distinct antenna, the transmitted signal based on the weighted signal copy. Illustratively, the first signal is a spread spectrum signal. Moreover, the process of weighting comprisesapplying a phase shift to a sub-segment. The invention further provides a receiving technique which comprises, for a copy of a received signal, despreading a segment of the received signal, demodulating a plurality of sub-segments of the despreadsignal segment, wherein each sub-segment is demodulated with use of one or more estimated communication channel characteristics corresponding to the sub-segment, and forming a summation signal reflecting a summation of a plurality of demodulated sub-segments. The receiving technique further comprises forming a signal reflecting a signal segment value, the formed signal based on one or moresummation signals. The step of demodulating may precede the step of despreading in some embodiments.

Description

7 ~ 9 DIVERSITY FOR DIRECT-SEQUENCE
SPREAD S3PECTRIJI\I SYSTEMS

Field of the In~c..t;o..
The present Ll~,nlion relates generally to Direct Se~quence Spread S Spectrum wireless co~ ni~Pti-)n systems, such as Direct-Sequence Code Division Muldple Access systems.

Back~round of the I~
In cellular radio systems, each cell is a local geographic region c~ ui.~ g a base station and a plurality of mobile users. Each mobile user 0 C~ h S direcdy with a base station only; there is no direct mobile-~o-mobile C~ n The base station p~ . r .. ~, among other things, a relay funcdon allowing a mobile user to co.. ni~tP with a user in another location. So, for e~ mrle, the base statdon provides courlinE of a mobile user's llnn.~ l() to ~ c another mobile user in the same cell7 to another base station for coupling to a mobile 15 user in another celL or to an ordinary public switched telephone network. In this way, a mobile user can send and receive in rO. ~~~ n to and from any other addressable user.
Direct Sequence Spread Spectrum (DSSS) systems, such as Direct Sequence Code Division Multiple Access (DS-CDMA) systems, are attracting 20 wi(lespre~l attention in the personal co-----~ ti-~n fields, such as, for eY~mp~
digital cellular radio. In a DS-CDMA cc.. --i~ ~ion system, both the time and L~ u~,ney domains may be shared by all users sim lllt~n~on~ly (this ~imnlt~nl~ollc - sharing of time and requency domains is to be ~lic~ingllicl~ed from time-division and frequency-division multiple access systems, TDMA and FDMA, where multiple user 2s co...." ..~ ti~n is f~ it~tçd with use of unique time slots or frequency bands, respectively, for each user~. As such, a base station may simult~nf~ously transmit distinct inro....~licln signals to separate users using a single band of freqllen~ s.
Individual inro~ i( n signals sim~ neoll~ly lln~ lrd may be isolated by each receiving user because of ~e base station's ~ltili7:lti~)n of unique signature sequences 30 in the ll~n~ n of the infr. rm~tion signals. Prior to tr~n~m~ nn~ the base station multirli~s each infnrm:-tinn signal by a signature sequence signal assigned to the user intended to receive the signal. To recover a tr~n~mitt~d signal from among those signals tr~n~mined cim-llt~n~QUsly in a frequency band, a receiving mobileuser mnltipli~s a received signal (co~ g all LIA~ lrd signals~ by its own - 2 - ~ r~

unique signature sequence signal and integrates the result. By so doing, the user rntifif~s that signal intended for it, as distinct from other signals intended for other users.
In wireless cc.""".~ Ation systems ~such as DS-CDMA systems), an S in~l " ",l ion signal is co" "- ,l,--irPt~d from a l~ lilirl to a receiver via a channel comrrising several lndependent paths. These paths are refelTed to as multipaths.Each multipath le~JIGSelll t a distinct route an h~ rO, ~ .l ir~n signal may take in traveling between tra-ncmitter and receiver. An i~Ço",.~iion signal col.. l-inAtf,d via such routs or multipaths appears at a receiver as a plurality of multipath signals, one 10 signal for each mnltirAth The Amrlit~l(les and phases of signals received from a lln~
through different mnltir:lthc of a co~ ir~tion channel are generally in~leFen(lent ~ -of each other. Because of complex addition of mnltir~th signals, the strength ofreceived signals may vary between very small and mrUlf,'rAtely large values. The5 phf flO~r ll~n of received signal strength v~iation due to complex addition of ~ ~-mnltirAth signals is known as fuding. In a fading e.lVIlUnll.~ , points of very low signal strength, or deep fades, are sepAr~tecl by al,~,.oAi-lla~ly one-half wavelength from each other.
~Illtir~th~ f ~~col~t~ .,,d in wireless cGll~ atir~n systems can be 20 ~s- ihed by cer~ain cll~n~ t~-;ctirs~ such.as ~mrlitllr1e ~tten~ tir)n and phase shifting. For ~YAmrle, the m~lltip ~thc of a DS-CDMA channel may provide different ~,,,pl;~ le At~emlAtionc and phase shifts to an i~ro~ Atinn sign~ co.~ rAt~d from a L~ . . .; l h" to a receiver. These different ~ . . .l.l i l . .-le and phase ch5 ~ n~ b ~ s may vary due ~o, e.g., relative ll,o~ t between tran~rnitter and receiver, or changes in 25 local ~,vO~ dp~ of the ~ or receiver due to ~llo~ . Because of the v~iation of mllltir~th chArnrten~ti( s. a receiver can ~ re a signal which fades witb time. This fading is a .n~.~ire,~ n of the complex addition of mllltip~th signals having time ~~arying A-mrlit~ldes and phases.
If the cha~ -t~ . ;.ctins of a DS-CDMA mllltirath vaTy slowly, a receiver 30 c ~ ei~ a deep ~ade may observe a weak signal for a long peAod of time. Long fades are not Imt~o~nmnn in, e.g., indoor radio systems, where relative movementbetween l~;cei~ and ~ is slow or non~Yi~Pnt (often, one of these two is an immobile base station; the other is a mobile device canied by a person). Since the duration of a deep fade may be large in u~ o~ to the duration of 3s i.lrOI " ,~l ion symbols being co. . " "~ ~ ,);r5~te~1, long bursts of symbol eIrors may occur (due to the weakness of received signal strength for an extended period of time)~

.. . . .
To avoid or mitigate the detrimental effects of fading, a technique providing diversity may be employed. Diversity refers generally to the ability of a col,llll,l,lir,:ltion system to receive infc-rmS~tinn via several in(llopçn/1çntly fading ch~nn~lc As a general matter, diversity techniques enhance a system receiver's S ability to combine or select (or both) signals arriving from these independently fading chSlnnç1c thus enabling (or fs~nilit~ting) the extraction of co""ll~ ed illr~ ion.
: ~ ' SUI~ ofthe ~Yention : ~:
The present invention provides a t~hni~lue for mitig~ting ~e 10 de~ l effectsoffadinginDSSS systems. An illustrative h~nsmittçr ' embodiment of the n~ ,nlioll provides diversity by inilu~ucillg a sr4~.ence of *istinct weights to SCg~ i of a signal to be tri~n~mi~tç~ Spe~ific~lly, given a signal to be ll;1Ir,d which col~ es si~nal segments Teflecting binary digits, the illu~l a~
embodiment forms M copies of the signal, wheTe M is the number of antennas used ~ ;
15 in I l ,.n~l il;l 1 ;I~g the signal. For each segment of each signal copy, the embodiment of the invention applies a distinct phase shift to each of M sub-se~ll,. .-1~ of the : .
segment. As a result, M phase-shifted signal copies are produced, one copy for each antenna. The se4u~,nce of distinct phase shifts applied to a given segment copy is itself d;stinct from the sequences of phase shifts app]lied to any other copy of the~0 given segm~nt Each of the M phase-sllirl~,d signal c:opies forms the basis of a signal r.d to a receiver with use of a distinct antenna.
An illu~L(alivri receiver çmho~limçn~ CC~ es a plurality of receiver branches, each branch coll-,~onding to a mll1tir~th of the commllnir~tir~n channel through which ~ d signals have been sent. Each receiver branch pe.ru~ s 2s despreading and (3emn~ tion p ucesses. The despreading process eon.~ ,s forming a product of a segment of the received signal and a signature sequence signal. 'Values of the despread received signal COllc~ n-lh-~ to a sub-segment are summed. The resul~ing sum is provided to a (1çmf)(~ll1~ 'on process which operates to remove the e~fects of the m~lltip~th on received signal amplifn ~~ and phase. Because 30 of the distinct phase shifts applied by the ~ to sub-segments of each signal segment reflecting a binary digit, the clemr ~3Ul:~ti~n process operates on a sub-segment by sub-segment basis. The demnd~ t~d sub-segment values for each segrnent are sllmm~d Summed sub-segment values from each receiver branch form the basis of a ~I~t~.rmin~tion of the value of the binary digit corresponding the 35 segment in question.

~4~ 2~ 7~

Illu~LId~ivt; embo~im~ntg of the invention provide diversity of ML'th order, where M is the number of antennas employed by the tr.qnemitter and L' is the number of receiver branches co~ ,onding to L' ml-ltip~thg While an il~ alivt;
receiver embo~ of the invention may incol~lale a multi-branch RAKE
s receiver (as eYrlqin~d below), it wvill be apparent to those of ordinary sldll in the art that a receiver with but one branch (i.e., L' = 1) may . lso be used. Though illustrative embotlimPnte of the present invention concern DS-CDMA systems, the present invention is qrplir,q,h'~ to indoor and outdoor DSSS systems gener. lly, such as DS-Carrier Sense Mul~iple Access systems, etc. Therefore, the invention has 10 applicability to cellular telephony, wireless PBXs, wireless LANs, etc., and may be used in c- mhinqtinn with other DSSS systems to enhance diversity.

Brief Dcs~ liul~ of the D~
Figure 1 presents informqtinn, signature sequence, and spread spectrum signals illus~laliv~ of DS-CDMA ~"..,~,..;e~
Pigure 2 presents two signal phasers from two tr-nemitting antennas at specific points in space where deep fades occur.
Figure 3(a) and (b) present fliqgrqme relating to received signal m~nitnde wi~hout and with, lesyeclively, the operation of a n e~l-bodime.lt the invention.
Figure 4 presents an illu~Llalive DS-CDMA l~n~",~ in accordance with the invention.
Figure S presents an illu;~ iivt; DS-CDMA receiver hl acc~Jldallce with the invention.
Figure 6 presents a s~u~,nce of complex mllltir~th char~teri~ ti~s ~or 2s use by the illu~ live receiYer ~ ent d in Figure 5.
Figure 7 presents an illustrative ~rr,.~,."i~ encoder for use with emboflim~nte of the invention.
Figure 8 presents an illu~llati~v DPSK clemn~ tnr for use with the receiver presented in Figure 5.

'"''':

-5- 2~ 7~

Detailed Description A. Introduction The illustrative embodiment of the present invention concerns a wireless :
DS-CDMA co~ ic~tion system such as, e.g., an indoor radio co"""~ rs~ n 5 system, a wireless local area network, a cellular te~lephone system, or personal cGIllllllll~ir~tir~ns system. In such systems, a base station commnnly uses a plurality of antennas (e.g., two) for receiving signals ~ led by one or more mobile units. This plurality of antennas provides the base station with a form of diversity known as space diversity. In accc,ldallce with the present invention, a plurality of lo antennas at the base station should be used for the tr~n~micsion of signals to mobile units. Advantageously, the same plurality of antennas used for base station reception may be used for tr~n~mic~ n to the mobile units. These mobile units need employ but one antenna.

1. DS-CDMA Signals Figure 1 presents a basic set of signals illui,l~ative of DS-CDMA
tr~-n~mi~iion. Signal a(n) of Figure l(b) is a signature sequence signal ~c~
with a particular receiver, as ~;icl~sed above. Signal a(n) comprises a series of rect:lng~ r pulses (or chips) of duration Tc and of mslgnitl~(le i 1. Discrete time variable n indexes Tc intervals (i.e., n is a sampling time at the chip rate).
Signal b(n) of Figure lta) is a signal (e.g., an i,lr." ,.~lit)n signal) to be c~-"~"lll";c:~ted to a receiver. Each bit of signal b(n~ is of a duration T and is indexed by i. As shown in Figure l(b), there are N chip intervals of duration Tc in inteîval T (i.e., N=T/Tc).
Tlle product of these two signals, a(n)b(n), is a spread spectrum signal 25 pl~enLt;d in ~Figure l(c). As shown in Figllre l(c), the first N chips of the spread spe.;~,ulll signal are the same as the first N chips of signal a(n). This is because signal b (n) -- 1, û S n < N -1. Moreover, the second N chips of the spread ~ ~ Ll Ulll signal have polarity oRosite to that of the second N chips of signal a(n), sincesignal b(n)=- 1, NSn~2N~ us, signal b(n) mod~ tes signal a(n) in the 30 classic sense.

- 6~ 7 ~ 9 2. Fading in DS-CDMA Systems Figure 2 presents an indoor radio sys~em comrri~in~ a base station 1 having two antennas, Tl and T2, for ~ ,s.,~ ;n~ a signal through, for eY:3mrle aRayleigh fading channel to a mobile receiver (a Rayleigh fading channel is a channel S without a line-of-sight path between ~ and receiver). Each of the antennas Tl and T2 tran~mit~ a spread spectrum signal, u(n), which reflects a scaled product -of signals a(n) and b(n) shown in Figure l(c). Each copy of signal u(n) nl~es an inrl~pPn(lPnt change in ~mr1itl~ and phase due to the mllltir~lth in which it travels. This change in ~mrlitll~lP and phase to the ~ d signal is lo e~ sedasacomplexfading coeffil~iPnt"B1 (n),where1, 1~lSL itl~.ntifil~s the mllltir~th (in Figure 2, L=2).
The signal received by receiver R I, s(n), reflects a sllmm~ n of the a~ rd signals L
s(n) = ~,A,BIa(n-~l)b(n-~ v(n) (1) 1=1 15 wheTe A is a ll.. "~ , gain factor, ~1 is a ~ ;oll delay s~c~oci~ted with a particular mllltir~lth, and v(n) is C}~n~si~n noise added by the channel. Signal s(n) lh~,lcrcllc comrriies a sllmm~ti()n of received signal phasors Sl, where Sl ~ A,Bla(n-~l)b(n-~l). , . .
In the example of Figure 2, signals S 1 and S 2. are received at specific 20 points in space where a deep fade occurs. The deep fade is due to a destructive r~ e of S I and S2. Signals S 1 and S2 are independendy and i~l~ntir~lly distributed with, e.g., Rayleigh ~mrlituc~e and uniforrn phase. The complex fading ç7. .~ Ir~ s of the channel through which phasors S 1 and S2 are cc ...l~ .ir~tPd (~B l (n) and ~B 2 (n)) change slowly, so that the deep fade eY~ ~ced by receiver R
25 of Figure 2 is e~çnti~lly static.
The deep fade shown location (b) of Figure 2 occurs because of the weakness of received signal energy from each individual antenna T 1 and T2. Thus, despite the fact that received signal phasors are not destructively aligned, receiver R2 eYreriPn~es a fade.

30 3. Path Diversity in Cv~ ..al DS-CDMA Systems Among the techniques used to mitigate the effects of fading in DS~
CDMA cc~ on systems is the path diversity technique. Path diversity in DS-CDMA systems entails estim:~tion of the deiay introduced by each of one'or ~7~ 2:~a~7~

more mllltir~hc (in comparison with some reference, such as line-of-sight delay), and using this delay in a receiver structllre to separate (or resolve) the received mnltir~th signals. Once separated, conventional techniques may be used to selectthe best mllltirath signal (or to combine mllltil~th signals) so as to extract the S cc~-"--....-irptt~.d i..r(,...-~lion A receiver structure often employed to provide path diversity is the so-called RAKE receiver, well known in the art. See, e.g., R. Price and P. E. Green, Jr., A Communication Technique for Multipath Channels, 46 Proc. Inst. Rad. Eng.
555-70 (Mar 1958).
While the path diversity afforded by conventional RAKE receivers is bel~c,rlcial in many i~ es. it may not provide a gi~nifi~On~ diversity benefit in certain ci,~ rec, such as some indoor radio el~vil~ .nt~ This is because the range of mnltir:lth delay values in these ellvu(J-Illlel~ls is small (on the order of 200 to 300 nanoseconds) ~OIl~pal~d with the duration of a DS-CI)MA chip interval 15 (which may be, for eY:~mrlP~ s). Because of this, knowledge of delay values is ininmripnt to allow resolntion of multipath signals. Thus path diversity is not generally available in such con~e.~lional DS-CDMA systems.

4. Inl~ to~heIllu~t dti~: Embo~ J~
The illuDhali~e el-lbo~3i .~ ; of the present invention provide diversity 20 for a DS-CDMA systems, even in indoor radio environmPnt~
The illusLI~llive trsln~mitter embodiment of the present invention introducesphaseshifts~l(n)and~2(n)tosignalsul(n)andu2(n)llon~---il1Pd from antennas Tl and T2"~,D,.,e~;lively. These phase shifts are introduced for aportion of the interval T co.l~,~on(lillg to each il-f.l. " ~lion signal bit. ~ese phase 2s shifts have the effect of reposih(-nin~ signal phasors S I and S-2 with respect to each other. Should the signal phasors be disposcd such that they add destructively, the phase shifts work to alter signal phasor angle so that the signal phasors add co~ luc~ivt;ly. This con~lluclive addition mitig~tes the effects of fading.
It will be understood by those of ordinary skill in the art that the relative 30 angular position of signals S 1 and S 2 in Figure 2 is merely illu~ ivt; of the possible relative angular positions such signals may take. However, signals S 1 and S 2, being out of phase by nearly ~ radians, ~ enl a near worst case scenario.
Since the operation of the embodiment of the present illvt;l~ works to mitigate worst case scenarios, less severe cases are naturally accounted for by the 35 embodiment.

,~ . , ~ .

-8- 21~97~9 The operation of the illustrative tr~ncmi~er embodiment may be further understood Wit}l reference to Figures 3(a) and (b). Figure 3(a) presents signals S l and S 2 as they appear in Figure 2. As a consequence of the static angular oriçnt~tis~n of these signals, the resultant sum of these phasors, s (n), has a ms~nitucle, S ¦s(n)¦ = Gl, which is small compared with the m:lgnitu(lP of the individual signals.
nitu~Ae G I is indicative of a deep fade. ~ssuminp no changes in these signals, a given illÇo,.,.,.l;cm bit, such as bit b(n), 0SnSN- 1 (and likely many more), would not likely be received.
In Figure 3(b), a phase shift of ~ radians has been applied by the l0 llOn~ l to signal u 2 during the first half of the bit interval (i.e., O ~ n ~N/2 - l) in acc.,~ ce with the invention. This phase shift has the effect of çh~n~ine therelative angular disposition of S 1 and S2 such that the destructive intvlrt;~ cv e~ by reseiver R 1 becomes constructive. As shown in the Figure, the m~gnitu-leofthesumofthephasors,ls(n)lisG2 forthe-firsthalfoftheintervaland 15 G 1 for the second half of the interval. The large m~nituc~P, G2 for a portion (or time segment) of the bit interval enables the bit to be received by receiver R l .
The illu~ Liv~ l.,..-.~...;li.., embodiment may be e~tPn(l~d to deal with the deep fades shown at location (b) of Figure 2. All that is required is the use of 1ition:~1 tr~ncmittine antennas to help contribute to received signal strength. A
20 11iscucsion of the embodiment below is generic to the nurnber of tran.~mittin~
~ntl.nn~, M.

5. F ~ o.1;~ -' Ha~ .dfe For clarity of e= ~ n~l ion~ the illu~llalivr vnll~odil.,vnl of the present iOII is ~lvsvnlvd as comrrising individual function:ll blocks (inr~ inE~
2s fnn~ti~n~1 blocks labeled as "I lUCvssv-~"). The functions ~hese blocks lv~lvse.ll may be provided through the use of either shared or de~irated hardware, inc1u-1in~, but not lirnited to, ha~d~ v capable of eYec-lting software. For eY~mr1P, the functions of ~lUCvSSvl~ presented in Figare S may be provided by a single shared ~ ces;~Or.
(Use of the tenn "processor" should not be construed to refer exclusively to hald~. rAlG
30 capable of e.Y~cutin~ software.) Illustrative embo-lim~nts may co"~ e digital signal processor (DSP) ha..lwd,c, such as the AT&T DSPI 6 or DSP32C, read-only memory (ROM) for storing software pelru.. -Ji--g the operations ~ cussed below, and random access ~ ;
memory (E~AM~ for storing DSP results. Very large scale integration (VLSI) 3s h~dwal~; embo~im~nt~ as well as custom VLSI circuitry in combination with a ;

- 9 - 2 ~ 7 ~ ~

general purpose DSP circuit, may also be provided.

B. An IllustrativeTr~ r Ell.bo~ nt Figure 4 presents an illul7~allvt; 1l0~ ,- embodiment in acco,dance with the present invention. The ~ ,- receives a signal, b(n), for tr~rlemiesion 5 to a receiver. Signal b(n) is "spread" in the conventional sense of DS-CDMA
systems by multiplying the signal by a signature sequence, a(n), provided by signal UI 12. This m~lltipli- ~tion is pf,- rO, ",~ tl by multiplier circuit 10. The result of this multiplica~ion is a spread SIJ~IIU111 signal reflecting the product a(n) b(n).
This spread .l~ecLIunl signal is provided in parallel to a plurality of M 1l,."~";~
10 circuit antenna ~ nrh~o.s, Each such antenna branch comrri~e~ a mllltiplif~r circuit 15, a signal t,~ c~ 17, conventional 1 ~ n~ circuit~y 20, and an antenna 25.
The mllltirli~.r circuit 15 of each antenna branch applies to the spread ' ~e~ill Ulll signal (or weighs the spread s~e~ un- signal by) a distinct time-varying signal Pm (n) of the for n Pm (n) = Am (n) ej'~m(n) (2) where m indexes the antenna branch, Am (n) is signal :~mplituf'l~, and ~3m (n) is signal phase. Signal p m (n) is generated by signal gcne~ Of 17. ~mplinlde A m (n) of signal Pm (n) takes the form Am(n) = ~ . (3) 20 Phase ~ m (n) of signal p m (n) takes the form ~3 ( ) 27~(m- 1 ) m' (4) , where m indexes the antenna branch; and rn' = 1, 2, ..., M indexes equal temporal portions (or sub-segment:~) of a segment of the spread s~,e~ Ulll signal. Each such segment is an interval of length T and is associated with a bit of b(n). The equal 25 sub-se~n~nt~ are given by iN + (m'-1~ M ~ n SiN + m' M - 1, (5) , where i indexes the bits represented by signal b~n). The i11u~h~live embodiment therèfore applies a distinct phase shift, Qm (n) and a common gain Am (n) to each sub-segment of the spread sl~e~ ul.~ signal ~i~ociS~ted with a bit of b(n). If N is not ~ . . .. ~, .. ... ... .. ...... . . . . .. . . .. ... ......... .... ..

lo- 2 1 ~ 9 an integer multiple of M, the length of the sub-segments should be made as equal as possible.
The application of phase shift ~3m (n) by the operation of gellr.~ 17 and multiplier circuit 15 is illustrated with ~ef~l~nce to Figure 3(b). As discussed S above, when M = 2 a phase shift of ~ radians is applied to one of two 11,l n ~. "i l ir.d phaso~s d~ing the first half (m' = 1) of a bit interval. Given two tr~ncmittin~
antennas (i.e., M=2), the phase shift of 11 radians applied to the spread ~e~ ulll signal in the second antenna branch is provided by ~*..f..,.lr.. 17 in acco~ ce with expression (4). So, for example, the phase of P m (n), ~ m (n), is equal to 7~ when lo M=2, intiit~tin~ ~e two antenna 'o.,.,~l PS, m=2, int1it'~tin~ the second of the two cllei~ and m~ int1i~sltin~ the first of M=2 equal sub-segmPntg~ "
~e.1~ u! 17 applies phase shift ~m(n) for sub-segments defined in terms of n by e~ si~n (5). So, ~or eY~mrl~ s~lmin~ i= 0 (i.e., ~ ming the first bit of b(n)), and sl~bstitlltin~ M=2, m' = 1, and m=2, expression (5) gimrlifi~s to N

15 O~n~ 2 ~ 1 -- the first half (or sub-segrnent) of the interval c~~ pon~l;..~ to the first bit of b(n). Thus, ~,e~ ol 17 provides Pm (n~ with phase shift 9m (n) =~ for the sub-se~ ,nl defined by O~nS N ~
Ce~ . 17 operates in accol~.e~; with e~ o~C (4) and (5) to apply a phase shift of zero to t~e spread spe~ uln signal in the second antenna 20 branch during the second half (m' = 2) of the interval col.~ ding to the first bit of b (n). This zero phase shift is shown in Figure 3(b) by the phasor S 2 in its original position (shown in Figure 3(a)~. Moreover, ~ ltll 17 applies a phase shift of zero to the spread specllulll signal in the first antenna branch during both the first and second halfs (i.e., both sub-segn~.nt~) of the interval ccrrespontlin~ to the first bit of 25 b(n). Again, this is done in acc(llddnce with exp~essions (4) and (S). This zero phase shift is shown in Figure 3(b) by phasor S I lrl l l~ in its original position ;
(shown in Figure 3(a)) for both halves of the bit interval.
The distinct weigh~ing signals, P m (n), applied to each sub-segment of a bit interval (or segment) co.,~ le a sequence of weighting signals. The se.luellce of 30 weighting signals applied by one antenna branch of the elllbo lill~n~ for a given bit interval is distinct from the sequence of weighting signals applied in any otherbranch of the embodiment during the same bit interval. So, for example, the sequence of phase shifts applied by the first antenna branch of Figure 4 for thesegments of the bit interval di~cl~sed above is (0 rad., 0 rad.). This sequence is 35 distinct ~rom the sequence (7~ rad., 0 rad.) applied by the second antenna branch for -Il- 21 ~7~

the sub-segments of the same bit interval, since the first phase shift of each sequence is not the same.
The product of spread spectrum signal a(n) b (n) and signal Am(n)ei~m(n) producedbymultipliercir~uit lS of eachantennabranch lSm<Mis S a signal um (n). Each signal um (n) is provided to co~ iol~al tr:mcmiicinn circuitry 20. Circuitry 20 provides, inter alia, pulse- shaping, RF-mtul~ tion, and power amp1ific~tion in prep~r~tion for signal trangmigsinn via antenna 25.
As a result of the operation of the illu~lldlivt; Llo~ t. ~ lbo~li.-le.ll, each of M antennas 25 transmits a signal to a receiver. Each such signal is based on l0 a di~lh~ ly phase shifted version of a spread spectrum signal.
It will be understood by those of ordinary sldll in the art that a 1lo~ bodi~ in accordance with the invention may be realized with any number of antenna branches. Expressions (2)-(5) above are ~ ,sc.1ted generally to allow for such extended re~1ir:ltiong. Ful ~1lGII~ lG, it will be understood that the 15 sequence of operations which co~ ilule des~l~,adillg, as well as the s~uellc~, of despreading and ~1em~111SIti--n op~tiong, is illustrative. Other sequences of such operations may be realized in accol.lance with the present invention.

C. An Illustrative Receiver Fmho.~
Figure 5 presents an illu~lldlive DS-CDMA RAKE receiver embodiment 20 of the invention. The ~ bodi,lle1lt comrr1ces an antenna 50; conventional receiver circuitry 55; L' RAKl~ receiver branches, where L' ~1s less than or equal to thenumber of mll1tir~thc, L; s~lmmine circuit 80; and decision processor 85. The RAKE receiver branches are indexed by l, such that 1 ~lSL'SL. As is COI~v~lliOl~f~r RAKE-type l~C~ , each receiver branch is "tuned" to receive signals from a 25 particular m11ltir~th of a cs.. ~ ti- n channel. ;
The illustrative receiver elllbolilllel~l of Figure S may be used to receive signals ~ d by the illusl~a~ive 1~ emho~iml~.nt of Figure 4.
Acsnming M=2 transmi~ antennas and L' =2 RAKE receiver branches, use of the ' 1 illu;~lld~ivt; Ll~u~ and receiver in co,)-hi u ~ n provides ML'th (or in this case ' 30 fourth) order diversity.
Each RAKE receiver branch comrrices a DS-CDMA despreader 60, a demodulator 70, and a summ~tic-n memory 75. Receiver branch tuning is accomplished conventionally7 by estimation of m111tirs~h tr~ncmi~sion delay ~l (for use by despreader 60) and the complex conjugate of the m111tirath complex fading35 coefficient"BI (n) (for use by demorlu1~tor 70). Each despreader 60 comrri~ès - 12- 2 ~

multiplier circuit 62, signal generator 63, and summation processor 64.
DemQtlnl~tor 70 c-""l";~e~ demn~ ion l~lucesso~ 72.
Antenna 50 receives tr:~ncmitt~d ml-ltir~th signals from a tr~ncmittpr embodiment of the invention. The received signals, r(t), are processed by S conventional receiver circuitry 55 (comrricin~ e.g., low noise ~mrlifiPrs, RF/IF
band-pass filters, and a match filter) to produce signal s(n) as ~ cllcced above with ~GÇcl~,nce to expression (1). Signal s(n) is provided to each of the L' receiverbranches.
Multiplier circuit 62 receives signal s(n) from circuitry 5$ and a 10 delayed version of the signature sc~luence from signal ~,vn~ or 63. The signal gt~n,-~tors 63 of the embodiment are identical but for t~he delay they apply ~o the signature sequence. Each delay, ~1, is an estimate of the tr~ncmicsion delay ~Csoci~ted with the Ith mn1tir~th This delay is ~etermin~d by ~,e~ ,.ln. 63 in the conventional fashion for DS-CDMA systems. See, e.g., Pickholtz et al., Theory of15 Spread Spectrun. Communications--A Tutorial, Vol. COM-30, No. 5, ~E
Tr~nc~rtionc on Comrn. 855,870-75 (May 1982).
The ou~ut of multiplier 62 is provided to sllmm~ti- n l~lucessoi 64. For -each bit of signal b(n) to be r~ceived, l~lucei7i7ol 64 forms M sn~m~tionC of the signal s(n) a(n-~l ) provided by mnltirlier 62. Eac~h summ~ti(-n is of the folrn iN + m' N + ~
y~ s(n) a(n-~l) (6) n=iN + (m'-l) M + '~
where i refers to the ith bit of b(n), m' indexes equal length sub s~~ ,llL~ of the ith bit interval, and ~1 is the col.v~ ionally ~t~ ~illcd mlllhr~th tr~nsmissil n delay.
For each bit of b(n), ~l~ess(Ji 64 prûvides a despread signal segment which c~ esMoutputsignals, Ylm~, 15 m'<M.
Sû, for eY~mrle, if M=2, processor 64 will fûrm two su~m:~ti~nC~ each of which formed over one of the two (i.e., M) equal length sub-se~ ntc of the ith bit interval indexed by m'. These summ~innc have a forrn given by expression (6)~
iN+ N2 +~
Y~ s(n~a(n-~l) (7) n=iN+~cl . ~ .
iN+N+~
Yl2 = ~ s(n)a(n-~l) (8) n=iN+ 2 +~
, : ."
:, . .. ..... .... .... ... . .. . ..

- 13 - 2 ~ ~7~

Therefore, summation processor 64 trea~s the sub-segments of the ith bit interval separately, since such sub-segments are subject to distinct phase shifts applied by the tr~n~mitt~P,r.
The M output signals provided by sllmm~tion processor 64, YL~, for the S ith bit and the Ith mnlti~ th, are provided as input to (lemf)fllll~tion processor 72.
DPmndnl~ti-~n processor 72 multiplies each signal, Ylm'~ by an estimate of the conjugate of the complex fading coefficient for the Ith rml1tir~th In conventional RAKE receivers, the estimate of the conjugate of the complex fading coefficient for the 1th mllltir~th is determined on an in~;lGIlleJll~l bit by bit basis. That is, the 10 estimate of the conjugate of the fading coefficient for the ith bit is dependent on an estimate of the conjugate of the coerrlciellt for the i- 1th bit. However, because of the application of different phase shifts in different se~l,lr ll~ of the ith bit interval (bythe ni..l,lllill~,,), this ill~.;lGIll~ ll clplcl~ nslti~on oftheconjugateofthecomplex fading coefficient must be modified. This mnrlifir~tinn may be understood with 15 reference to Figure 6.
As shown in Figure 6, for the case where M = 2, each bit, e.g., i = 1, has ~soci~t~dwithittwocomplexfadingcoerr~ e.l~ ,Bm~(iN),N<m'SM(=2). The second of these two coeffici~P.nt~"B2~ (N), is not dependent on the coefficient which imm~ tP.ly precedes it"Bl~ (N), but ra~her on the second of the two fading 20 coçffiri. .nt~ associated with the preceding bit"B2~ (o). This is because both cOerrlc;~ ,B2~ (N) and ,B2~ (0), correspond to a bit interval sub-segment specified by m' = 2. Therefore, such coçffir;enti reflect the s~me phase shift applied by the The dependence of co-Pffic;~Pntg is in(lir~tP.d in the Figure by an arrow 25 comlcc~ g a later coerrlciel~l with an earlier coçr~ n As may be seen from the Figure, a coeffir,i~Pnt a ~01, :~ePd with a given sub-segment m' of a given bit is d~endent on the coçffi~;Pnt of the same sub-segment of the preceding bit.
Therefore, processor 72 may be realized with M conventional co~PMr;Pnt est;m~tirn phase-locked loops, each such loop c~ e ..r,d with the same sub-segment m' in 30 successive bit intervals. See, e g., Gitlin, et al., Data ~omm~1nir Itirnc Prinrirl~s, 403-32 ~1992~. It should be understood that ~luce~ 72 of the illustrative receiver need estimate coefficient phase only. This is because the illu~ ivt; I~
embodiment uses only a phase shift to ~rf~n~idle the signals ~ rd by the dif~erent antennas.

.. .
~ .. : : , . ..... . .

-14- 2~ 7~9 Referring again to Figure 5, the output of processor 72 of the 1 mlllti~ th receiver branch for the ith bit compTiges over time M signals of the form Zlm' = 1~1 Ylm' (9) where the M signals are indexed by m'. These M signals are stored by snmmstion S memo~y 75 and added together as received. Memory 75 forms a sum as follows:

z~ m~. (10) m~=l Signals zl from the memory 75 of each receiver branch are summed by s~lmmin~e circuit 80. The result is a signal zi which reflects each received bit i. Signal zi is providecl to a conventional decision processor 85, which assigns a binary value for 10 each bit, bi, basecl on zi. Processor 85 iliu~lld~ively provides a threshold ~3etection such that bi = 1 when zi 20, and bi = 0 when zi < o. Binary signal bi is thus the received bit stream.
The embotlimt~nt~ of the l".t.~ t~,~ ancl receiver presented above concern binary phase shift keying (BPSK) m~n1~tit~n formats. However, other 5 mnfl~ tion formats such as binary ~liLr~ ial phase shift keying ~PSK) may be used. The k,.~ embodiment ~ .Led above may be a~lgrnt~ntPd to provide DPSK mn l~ tinn by use of the conventional dirr~ ,n~ial encoder 100 presented in '~
Figure 7. For DPSK mr clll1~tion of a binary signal d(n), signal d(n) is presented to theconventionalmod s ~ "~i"gcircuitllOofdirr~.entialencoder100. Modulo-2-snmming circuit 110 also receives input from delay 120. The output of înod-2- , , summing cirruit 110 is provided to the 1l~ .., embodiment as signal b(n).
Signal b(n) is also fed back to the mod-s~lmming circuit via delay 120.
The il~ alive receiver can be mn~1;fipd to receive DPSK mn~llll~t~
signals from the tr~qn~m;ttPr by replacing d~mo-l-ll~tion ~llUCes~ol:~ 72 ~liccngged 25 above with the (lemn~ ti~ m ~lucesso~ 73 psesented in Figure 8. Each processor 73 is shown comp~iging a loop co",~ g delay 130, conjugate plvcessol 135 and summing circuit 140.
The segments of signals ~lice~ce~l above in the context of the illu~llalivt; embodiments of the present invention concern individual binary digits (or 30 bits~ of a digital signal. It will be understood by those skilled in the art that these signal segllle.lls may reflect values of other types ûf signals in other emborlipmntg of the present invention. For example, in such embo(limPntg these segments may reflect complex-valued signals, analog signals, discrete-valued signals, etc.

Claims (20)

Claims:

1. A method of operating a direct-sequence spread spectrum transmitter for communicating a first signal to a receiver, said transmitter including a plurality of M antennas, said first signal including one or more signal segments, the method comprising the steps of:
a. forming M copies of said first signal;
b. for a segment of a signal copy, weighting each of two or more sub-segments ofsaid segment with a distinct signal, wherein a sequence of said distinct weighting signals for said segment is distinct from sequences of signals weighting the same segment of one or more other signal copies; and c. for each of M weighted signal copies, transmitting a signal to said receiver using a distinct antenna, said transmitted signal based on said weighted signal copy.

2. The method of claim 1 wherein the step of weighting each sub-segment with a distinct signal comprises applying a distinct phase shift to the sub-segment.

3. The method of claim 2 wherein the phase shift for an m'th sub-segment, 1 ~ m' ~ M, of a segment of a signal copy which forms the basis for the transmitted signal from the mth antenna, l ~ m ~ M, comprises .

4. The method of claim 1 wherein the step of weighting each sub-segment with a distinct signal comprises applying a gain to the sub-segment, the gain comprising .

5. The method of claim 1 wherein the first signal is a spread spectrum signal which reflects a product of a second signal and a signature sequence signal.

6. The method of claim 5 wherein the second signal is a differentially encoded signal.

7. The method of claim 1 wherein a signal segment reflects a discrete signal value.

8. The method of claim 7 wherein the discrete signal value comprises a binary digit.

9. The method of claim 1 wherein the step of transmitting a signal to a receiver comprises forming a spread spectrum signal reflecting a product of the weighted signal copy and a signature sequence signal.

10. The method of claim 9 wherein the first signal is a differentially encoded signal.

11. A method of operating a direct-sequence spread spectrum receiver to determine one or more signal segment values which are represented in a received signal from a communication channel, the received signal reflecting signals transmitted with use of a plurality of antennas, the method comprising the steps of:
1. for one or more copies of the received signal, a. despreading a segment of the received signal, b. demodulating a plurality of sub-segments of the despread signal segment, wherein each sub-segment is demodulated with use of one or more estimated communication channel characteristics corresponding to the sub-segment, and c. forming a summation signal reflecting n summation of a plurality of demodulated sub-segments; and 2. forming a signal reflecting a signal segment value, said formed signal based on one or more summation signals.

12. The method of claim 11 wherein the step of despreading a segment of the received signal comprises the steps of:
i. applying a signature sequence signal to the received signal; and ii. for each of a plurality of received signal sub-segments to which a signaturesequence signal has been applied, forming a summation signal reflecting a summation of signal values of such sub-segment;
wherein a despread signal segment comprises a plurality of summation signals from step ii.

13. The method of claim 11 wherein the received signal is provided by receiver circuitry coupled to an antenna.

14. The method of claim 11 wherein a signal segment value comprises a discrete value.

15. The method of claim 14 wherein the discrete value comprises a binary digit.

16. A method of operating a direct sequence spread spectrum receiver to determine one or more signal segment values which are represented in a received signal from a communication channel, the received signal reflecting signals transmitted with use of a plurality of antennas, the method comprising the steps of:
1. for one or more copies of the received signal, a. demodulating a plurality of sub-segments of a segment of the received signal, wherein each sub-segment is demodulated with use of one or more estimated communication channel characteristics corresponding to the sub-segment, b. despreading a plurality of demodulated sub-segments of the received signal, and c. forming a summation signal reflecting a summation of a plurality of despread sub-segments; and 2. forming a signal reflecting a signal segment value, said formed signal based on one or more summation signals.

17. The method of claim 16 wherein the step of despreading a segment of the received signal comprises the steps of:
i. applying a signature sequence signal to a plurality of demodulated sub-segments; and ii. for each of a plurality of sub-segments to which a signature sequence signalhas been applied, forming a summation signal reflecting a summation of signal values of such sub-segment;
wherein a despread signal segment comprises a plurality of summation signals from step ii.

18. The method of claim 16 wherein the received signal is provided by receiver circuitry coupled to an antenna.

19. The method of claim 16 wherein a signal segment values reflect comprises a discrete value.

20. The method of claim 19 wherein the discrete value comprises a binary digit.