Ulllted States Patent [19]
`Gitlin et al.
`
`[54] SYSTEM AND METHOD FOR OPTIMIZING
`SPECTRAL EFFICIENCY USING TIME-
`FREQUENCY_CODE SLICING
`
`
`
`Inventors: Richard I)I Zygmunt Haas, Holmdel; Mark Silver;
`
`
`
`.
`
`Karol, Fair Haven; Clark Woodworth,
`Rumsofh all of N1
`
`US006018528A
`[11] Patent Number:
`[45] Date of Patent:
`
`6,018,528
`*J an. 25, 2000
`
`6/1993 Wagner .
`5,221,983
`5,260,967 11/1993 Schilling ............................... .. 375/203
`5,272,556 12/1993 Faulkner et al. .
`5,295,153
`3/1994 Gudmundson ........................ .. 375/205
`
`
`
`Flllgl'lllIIl 618.1. ..................... .. ll5eppanen .............................. .. 370/342
`
`,
`
`,
`
`575777024 11/1996 Malkamaki et aL
`5,581,548 12/1996 Ugland et al. ........................ .. 370/337
`
`en ................. ..
`
`..
`
`[73] Assignee: AT&T Corp, Middletown, NJ.
`
`[
`
`]
`
`Notice.
`
`pitent 1s subJect to a terminal dis
`
`Primary Examiner—Huy D- V11
`Attorney, Agent, or Firm—Jose R. de la Rosa
`[57]
`ABSTRACT
`
`[21] Appl.No.: 08/234,197
`_
`_
`Apr‘ 28’ 1994
`[22] Flled'
`[51]
`Int. c1.7 ...................................................... .. H04J 4/00
`[52] US. Cl. ........................ .. 370/436- 370/441- 370/468-
`370M752. 370/479’. 375/201’
`’
`3706;) 18 19
`[58] Field of Search
`329 3,35 336’
`370/20
`337’ 342’ 343’ 345’ 431’ 441’ 442’ 465’
`479’ 480’ 498’ 535’ 536’ 537’ 546’ 477’
`465;, 375,000’ 201’ 202’ 203’ 204’ 205’
`’
`’
`’
`’
`’
`’ 206’
`
`[56]
`
`References Cited
`
`US. PATENT DOCUMENTS
`37050
`9/1989 Suzuki
`4868 811
`37050
`4’914’649 M1990
`375/206
`5,029,180
`7/1991 Cowart ................... ..
`370/330
`5,134,615
`7/1992 Freeburg et a1.
`5,210,771
`5/1993 Schaeffer et a1. ..................... .. 375/203
`
`"""""
`
`A system and method for optimizing usage of a communi
`cations transmission medium. The transmission medium
`may be sliced into time and frequency domains so as to
`create time-frequency slices for assignment to users having
`varying access rates and user-application requirements
`Through Scheduling of the Various Speed users Within the
`frequency and time domains, the system and method can
`ef?ciently allocate and make use of the available spectrum,
`thereby accommodating higher rate users requiring greater
`bandWidths and time slot assignments While still preserving
`cost-efficient access for loWer speed users. Depending on the
`signal modulation scheme, the time-frequency slices may be
`allocated on non-contiguous frequency bands. The system
`and method is also applicable to code-division multiple
`access (CDMA) techniques by slicing the available code
`space along time-code domains, frequency-code domains or,
`in three dimensions, along time-frequency-code domains.
`Users may be efficiently scheduled based on code space
`requirements so as to optimize use of the communication
`medium
`
`15 Claims, 8 Drawing Sheets
`
`42
`
`F7
`
`F6
`
`F5
`
`F4
`
`40/ F3
`
`B
`
`c
`
`c
`
`D
`
`E
`
`F2
`
`F1
`
`F0
`
`F
`
`FREQUENCE
`BANDS
`
`A
`
`|
`
`J
`
`K
`
`H
`
`P
`
`S
`
`52
`
`M
`
`N
`
`0 /
`
`R
`
`J
`
`0
`
`[
`
`T \
`
`52
`
`/
`
`{100
`
`— CONTROL
`
`S0
`
`S1
`
`S2
`
`S3
`
`S4
`
`S5
`
`S6
`
`44
`TIME SLOTS _-_>
`12R HIGH-SPEED USERS: A,B,G,L
`5ORMEDIUM-SPEED USERS: C,E,F,H,I,J,M,0,0
`RLOW-SPEED USERS: D,K,N,P,R,S,T
`
`

`
`U.S. Patent
`
`Jan. 25,2000
`
`Sheet 1 0f 8
`
`6,018,528
`
`FIG. 1
`
`I
`\
`
`TIME TIME TIME
`TIME TIME TIME TIME TIME
`SLOT sLoT SLOT SLOT SLOT SLOT SLOT SLOT
`RAD|O12345678
`FREQUENCY!
`l
`I7‘I/I l
`l
`l
`I
`BAND \
`/
`/
`4/ w
`FRAME/\Z
`
`FIG. 2
`
`/ RF CHANNELS
`
`12 /FREQUENCY1 CONTROL CIRCUIT
`§\
`FREQUENCY 2 CONTROL CIRCUIT
`FREQUENCY \
`DOMAIN <
`I
`
`FREQUENCY3
`
`VOICE CIRCUIT
`
`‘0
`
`FRE UENCY 4
`Q
`
`VOICE CIRCUIT
`
`“
`FREQUENCY N
`
`‘'
`VOICE CIRCUIT
`
`TIME ——>
`INDEPENDENT
`
`BANDWIDTH OF EACH
`CHANNEL TYPICALLY
`LESSTHAN SOKHZ
`(IN EACH DIRECTION)
`14
`
`ALL VOICE CIRCUITS ARE
`FULLY TRUNKED, CONTINUOS
`TRANSMISSION CIRCUITS
`ONE CIRCUIT PER
`RF CHANNEL
`
`N
`
`

`
`U.S. Patent
`
`Jan. 25,2000
`
`Sheet 2 0f 8
`
`6,018,528
`
`FIG. 3
`
`20
`\
`REAL TIME
`
`22
`ONE FRAME
`
`22
`\
`FRAME
`
`FRAME
`
`FRAME
`
`FRAME
`
`FRAME
`
`:i/ //
`*5
`‘1S
`
`ms \\ \j}
`‘KS
`{g5
`
`6
`
`24\
`ONE TIME SLOT
`
`28\,//
`cg??g‘
`
`
`
`DATA X ns
`
`\\ D/2s
`Gill/mg‘
`
`UN'VERSALT'ME SLOT
`
`VOICE
`l
`l
`
`l
`
`n
`
`VIDEO
`|||llH||H|I|I|IIHIII||I|H|N||I||II|HH||||||IIHH|A
`
`n
`
`I
`
`|
`
`l
`
`n
`
`FIG. 4
`
`GUARD—-
`
`LOW-SPEED
`
`|-||G|-|-5PEED
`
`GUARD
`
`‘ BAND u HEADER U
`
`DATA
`
`“ BAND ‘
`
`)1
`
`28
`
`//
`
`= 26
`
`PACKET
`
`27
`
`>
`
`\\
`
`28
`
`4
`
`SLOT
`
`;
`
`

`
`U.S. Patent
`
`Jan. 25,2000
`
`Sheet 3 0f 8
`
`6,018,528
`
`/
`
`42
`
`FIG. 5
`/\
`
`F7
`
`F6
`
`F5
`
`F4
`
`40/ F3
`
`B
`
`0
`
`c
`
`D
`
`E
`
`F2
`F1
`
`F0
`
`F
`
`FREQUENCE
`BANDS
`
`A
`
`|
`
`J
`
`K
`
`\
`
`H
`
`P
`
`s
`
`52
`
`M
`
`N
`
`0 /
`
`R
`
`J\
`
`0
`
`L
`
`T \> 52
`/
`
`10o
`
`If
`CONTROL
`
`S0
`
`S1
`
`S2
`
`S3
`
`S4
`
`S5
`
`S6
`
`44
`TIME SLOTS
`46RHIGH-SPEED USERS: A,B,G,L
`ggxMEDlUM-SPEED USERS: C,E,F,H,I,J,M,0,0
`‘LOW-SPEED USERS: D,K,N,P,R,S,T
`
`\
`
`

`
`U.S. Patent
`
`Jan. 25,2000
`
`Sheet 4 0f 8
`
`6,018,528
`
`/
`
`42
`
`FIG. 6
`/L
`
`F7
`
`F6
`
`F5
`54\\
`F4
`
`40’
`
`D
`
`FREQUENCE
`BANDS
`F1
`E
`54\
`F0
`‘
`
`B
`
`c
`
`B
`
`\
`
`A
`
`|
`
`J
`
`K
`
`H
`
`P
`
`s
`
`52
`
`Q /
`
`J\
`
`T\
`
`/
`
`> 52
`
`M
`
`N
`
`R
`
`0
`
`L
`
`0
`
`F
`
`so
`
`s1
`
`s2
`
`s3
`
`s4
`
`s5
`
`S6
`
`44
`TIME SLOTS
`jg‘\H|0H-sPEEO USERS: A,B,G,L
`SORMEDIUM-SPEED USERS: C,E,F,H,|,J,M,O,Q
`RLOW-SPEED USERS: D,K,N,P,R,S,T
`
`\
`
`

`
`U.S. Patent
`
`Jan. 25,2000
`
`Sheet 5 0f 8
`
`6,018,528
`
`\
`
`,
`52
`
`/
`
`43
`
`FIG. 7
`/\
`
`07
`
`C6
`
`05
`
`c4
`
`40/ C3
`
`B
`
`0
`
`c
`
`D
`
`E
`
`()2
`
`01
`
`C0
`
`F
`
`CODE
`SPACE
`
`A
`
`|
`
`J
`
`K
`
`H
`
`P
`
`s
`
`M
`
`N
`
`0
`
`R
`
`0
`
`L
`
`J
`
`T
`
`so
`
`s1
`
`s2
`
`s5
`
`s4
`
`s5
`
`S6
`
`44 TIME SLOTS -—->
`jgRHIGWSPEED USERS: A,B,G,L
`50 ‘MEDIUM-SPEED USERS: C,E,F,H,|,J,M,0,Q
`\ R LOW-SPEED USERS: D,K,N,P,R,S,T
`
`

`
`U.S. Patent
`
`Jan. 25,2000
`
`Sheet 6 0f 8
`
`6,018,528
`
`/
`
`43
`
`FIG. 8
`/\
`
`07
`
`C6
`
`05
`
`c4
`
`40/ C;
`
`B
`
`c
`
`o
`
`D
`
`E
`
`(32
`
`01
`
`C0
`
`F
`
`CODE
`SPACE
`
`A
`
`l
`
`J
`
`K
`
`H
`
`P
`
`5
`
`M
`
`N
`
`0
`
`R
`
`0
`
`L
`
`J
`
`T
`
`\
`
`,
`52
`
`F1
`
`F2
`
`F3
`
`F0
`
`F6
`F5
`F4
`W
`FREQUENCY BANDS —» 42
`1g\“H|GH—SPEED USERS: A,B,G,L
`50 R MEDIUM-SPEED USERS: C,E,F,H,|,J,M,O,Q
`\ R LOW-SPEED USERS: D,K,N,P,R,S,T
`
`

`
`U.S. Patent
`
`Jan. 25,2000
`
`Sheet 7 0f 8
`
`6,018,528
`
`FIG. 9
`/\
`
`/
`
`\
`
`05
`
`A
`
`120
`
`cs H
`
`C8 P C85
`
`04 B
`
`C1D
`
`C1 K
`
`C10
`
`CIT
`
`CODE
`SPACE
`
`00 E
`
`00 L
`
`S0
`
`S1
`
`S2
`
`S3
`
`S4
`
`S5
`
`S6
`
`44 TIME SLOTS —>
`46RHIGH-SPEED USERS: A,B,G,L
`33R MEDIUM-SPEED USERS: C,E,F,H,|,J,M,O,Q
`\ R LOW-SPEED USERS: D,K,N,P,R,S,T
`
`

`
`U.S. Patent
`
`Jan. 25, 2000
`
`Sheet 8 0f 8
`
`6,018,528
`
`31
`
`TIME SLON
`
`ZRHIGWSPEED USERS: F,G,J
`50 RMEDIUM-SPEED USERS: A,B,C,E
`\\ LOW-SPEED USERS: D,H,|,K,L,
`
`

`
`1
`SYSTEM AND METHOD FOR OPTIMIZING
`SPECTRAL EFFICIENCY USING TIME
`FREQUENCY-CODE SLICING
`
`1. TECHNICAL FIELD
`
`The invention relates to a system and method for maxi
`miZing usage of a communications transmission medium,
`and more particularly, to a system and method for maximiZ
`ing usage of a communications transmission medium While
`preserving optimum access to the medium for users of
`differing access speeds and While maximiZing spectral use
`and bandWidth ef?ciencies.
`
`2. PROBLEM
`
`Many communication systems today, such as the Wireless,
`satellite, personal communications, and cellular communi
`cations systems, typically exhibit certain common require
`ments. For example, to maximiZe their ?exibility, these
`communications systems typically require a variety of
`access speeds in order to support differing applications. In
`order to be economically viable, the systems should also
`offer a generally loW-cost access for loWer-speed users.
`Lastly, the systems typically strive for a high degree of
`spectral ef?ciency in order to maximiZe usage of the par
`ticular communications transmission medium.
`As is knoWn, certain data transmission architectures have
`been developed in communications systems to allocate
`communication resources to individual users on their
`demand. Typically, these architectures ought to be structured
`to permit various users to utiliZe the resources in a fully
`shared communications system. Thus, the various architec
`tures are generically referred to as “multiple access” archi
`tectures.
`Referring to FIG. 1, one multiple access architecture for
`maximiZing usage of the communications transmission
`medium is commonly referred to as time-division multiple
`access (TDMA). As knoWn to those skilled in the art, in
`TDMA each carrier frequency 1 is severed into one or more
`time frames 2 having a plurality of individual time slots 4.
`Each of the time slots 4 is assigned to a user as an
`independent circuit. Information is transmitted by the user in
`short bursts during assigned or speci?ed time slots, With
`users being scheduled for access to the time slots 4 accord
`ing to their information transmission requirements. As Will
`be appreciated, hoWever, in pure TDMA architecture both
`higher-speed and loWer-speed users share a common com
`munications bandWidth, typically by assigning more time
`slots per frame to the higher-speed users. The draWback of
`this architecture is that high-rate access (high speed data
`bursts) is required even for loWer-speed users, Which
`increases the cost and complexity of the systems employed
`by those loWer-speed users.
`A second multiple access approach for structuring a
`communications transmission medium, as knoWn to those
`skilled in the art, is referred to as frequency-division mul
`tiple access (FDMA). A depiction of the FDMA approach is
`illustrated in FIG. 2. Unlike TDMA, the FDMA approach is
`independent of time. In FDMA, a number of individualiZed,
`narroWband channels 12 are used across the frequency
`domain (spectrum) 10. Rather than being partitioned into
`individualiZed time slots across the channel, in FDMA, one
`circuit 14 is assigned per channel 12 and, typically, users can
`access any one of the frequencies 12 in the frequency
`spectrum 10. A draWback of a pure FDMA architecture is
`that the maximum bandWidth available to an individual user
`is oftentimes limited, even if the particular user desires a
`
`10
`
`15
`
`25
`
`35
`
`45
`
`55
`
`65
`
`6,018,528
`
`2
`large peak bandWidth for only a short period of time. In
`order to access greater bandWidth, the user often has to
`utiliZe a plurality of transmitters that alloWs him to access
`several frequencies at the same time. This may add to the
`cost of the systems employed by those users. Moreover, as
`only a single user can occupy any given frequency, regard
`less of the time that the user Will occupy a frequency(ies) 12,
`the frequency spectrum 10 may not be fully utiliZed.
`Attempts have been made to support users having differ
`ing communication requirements in various of the afore
`mentioned communications systems. For instance, to sup
`port users of arbitrary access speeds and to retain loW-cost
`access for loW-speed users, a “Universal Time Slot”
`approach has been proposed by R. A. Thompson, J. J.
`HorenKamp, and G. D. Berglund (Ph0t0t0nic Switching of
`Universal T ime-Slots, XIII International SWitching Sympo
`sium Proceedings, Session C2 Paper 4, Stockholm, May
`1990). A depiction of the Universal Time Slot approach is
`found in FIG. 3. In the Universal Time Slot approach, each
`transmission frame 22 in real time 20 is separated into a
`plurality of individual time slots 24 of a set duration (for
`instance, X nanoseconds). The individual time slots 24 can
`transmit a given number of bits for voice (n bits) or video (m
`bits) transmissions, using different amounts of medium
`bandWidth. A so-called “data transparency” is created in
`each of the time slots, in that the signals in each time slot are
`typically generated and received asynchronously.
`Another attempt to maximiZe use of communication sys
`tems has been proposed by Zygmunt Haas and Richard D.
`Gitlin using a “Field Coding” technique (Optical Distribu
`ti0n Channel: An Almost-All Optical LAN Based On The
`Field Coding Technique, Journal of High-Speed NetWorks 1
`(1992), pp. 193-214). Field coding, typically used for opti
`cal transmissions, addresses the costly handicap of requiring
`an optical sWitching node to operate at the peak data
`transmission rate. Field coding separates the sWitching rate
`from the transmission rate by employing differing bit rates
`for the header (26) and data ?elds (27) of the optical packets
`(see FIG. 4). Guard bands 28 are used to separate individual
`user transmissions. Because the sWitching node performs
`only the sWitching operation and does not need to process
`the data portion of the packet, the sWitching node can
`operate at the loWer header rate, alloWing the faster rate data
`?eld to pass transparently through the sWitching node.
`In both of the proposed approaches, users are alloWed to
`transmit at their oWn desired rate during their assigned time
`slots. HoWever, While suitable for optical media Where
`bandWidth is abundant, these techniques are in fact spec
`trally inef?cient. In the cases of the previously mentioned
`communication systems (for instance, radio), the available
`communications transmission medium is quite limited and is
`often costly; there is typically only a limited amount of
`bandWidth available for access by users of the various
`communications systems. Thus, techniques that make ef?
`cient use of the transmission spectrum are necessary.
`
`3. SOLUTION
`These and other problems are addressed by a system and
`method for maximiZing complete usage of the communica
`tions transmission medium according to the invention. The
`system and method recogniZe that the transmission medium
`can be partitioned in frequency, time and code domains, and
`through optimum scheduling, user packing Within the over
`all frequency-time-code domain can be maximiZed in order
`to optimiZe spectral ef?ciency. The system and method also
`preserve a degree of inexpensive access for users With loWer
`access speed requirements.
`
`

`
`3
`In one embodiment of the system and method according
`to the invention, the transmission resource, partitioned into
`the “time-frequency” domain, is divided into a plurality of
`time-frequency “slices” that are allocated to users according
`to their various transmission requirements. For higher speed
`users, frequency slots are usually assigned contiguously in
`order to optimiZe the design of modulation and transmission
`architectures (eg a single transmitter for higher rate users).
`In a variant of this embodiment, Where frequency adjacency
`requirements can be eased, higher speed users can be
`assigned tWo or more non-contiguous time-frequency slices
`to further maximiZe spectral efficiency.
`In a further application of the system and method accord
`ing to the invention, the time-frequency slicing approach can
`also be applied to data transmissions With code division
`multiple access (CDMA) to account for optimum packing of
`code space. The CDMA transmission spectrum can be
`partitioned into the code-time domains, code-frequency
`domains, or, in a three-dimensional approach, into the
`code-time-frequency domains so as to optimiZe use of the
`available code space.
`The system and method provide better spectral use than,
`for example, a Universal-Time-Slot approach, coupled With
`the ability to accommodate a Wide range of access rates, the
`provision of loW-cost end points for loW-speed users, and the
`need for only a single transmitter-receiver pair per user.
`
`4. BRIEF DESCRIPTION OF THE DRAWINGS
`
`FIG. 1 illustrates a TDMA multi-access architecture for
`structuring user access for a given band in the frequency
`spectrum;
`FIG. 2 depicts an FDMA multi-access architecture for
`structuring user access in the frequency spectrum;
`FIG. 3 illustrates a Universal Time Slot approach in
`communications systems;
`FIG. 4 illustrates a Field Coding approach in optical
`transmissions by varying header and data ?elds;
`FIG. 5 depicts one embodiment of a time-frequency sliced
`system in accordance With the system and method of the
`invention;
`FIG. 6 depicts a second embodiment of a time-frequency
`sliced system for non-contiguous time-frequency assign
`ments in accordance With the system and method of the
`invention;
`FIG. 7 depicts an embodiment of the system and method
`of the invention for use With time-code slicing in Code
`Division Multiple Access (CDMA) systems;
`FIG. 8 depicts one embodiment of the system and method
`of the invention for use With frequency-code slicing in
`CDMA systems;
`FIG. 9 depicts reuse of code assignments in time-code
`slicing in accordance With the system and method of inven
`tion; and
`FIG. 10 depicts a further embodiment of the system and
`method of the invention for use With time-frequency-code
`slicing.
`
`5. DETAILED DESCRIPTION OF THE
`INVENTION
`
`Turning noW to the draWings, Wherein like numerals
`depict like components, FIG. 5 illustrates a time-frequency
`slicing approach according to one embodiment of the inven
`tion. As illustrated, the overall time-frequency spectrum (or
`medium) 40 can be partitioned in both the time and fre
`
`65
`
`6,018,528
`
`4
`quency domains as a plurality of frequency bands (“slices”)
`42 (F0, F1, .
`.
`. FN) extending over a plurality of individual
`time slots (“slices”) 44 (S0, S1, .
`.
`. SN). For purposes of
`illustration and not of limitation, users of the spectrum can
`be categoriZed into three general groups: high speed users 46
`(here A, B, G, L); medium speed users 48 (here, C, E, F, H,
`I, J, M, 0, Q); and loW speed users 50 (here, D, K, N, P, R,
`S, T). As illustrated in FIG. 5, a plurality of time-frequency
`“slices” 52 are gridded into the overall time-frequency
`spectrum 40.
`In accordance With the system and method of the
`invention, it is assumed that all of the various signals
`transmitted by users 46, 48, 50 Will occupy at least one
`frequency band 42. Moreover, it Will be realiZed that due to
`the nature of the equipment typically employed by higher
`speed users 46, the high-speed users 46 Will have the ability
`to modulate their signals so as to cover one or more
`frequency bands 42. Thus, as depicted, the overall medium
`can be sliced so that loW-speed users 50 Will be permitted to
`?ll one or more of the available time slots 44 in a frame,
`While higher-speed users can ?ll one or more of the available
`frequency bands 42 or time slots 44.
`A further assumption is that one “unit” of “slice”, Which
`is taken to be one frequency band allocation for one time slot
`allocation, is the minimum amount of communications
`resource Which Will be available to a user. Unlike other
`transmission techniques (such as the Universal Time Slot
`approach of FIG. 3) no guard bands “28 ” are necessary
`betWeen contiguous frequency bands 42 or time slots 44, or
`both, that are allocated to a given user, thus optimiZing full
`use of the medium (realiZing, of course, that guard bands 28
`may be needed to separate different users). Where a single
`user occupies contiguous allocations, a continuous fre
`quency band 42 and/or a continuous time allocation 44 can
`be realiZed because that same user may utiliZe the space
`Which Would be normally occupied by guard bands 28.
`Examples of the unit slice are depicted in FIG. 5 by the
`time-frequency slice occupied, for example, by various loW
`speed users 50 (i.e., users D, K, N, etc.).
`Thus, through use of their respective transmitters (not
`shoWn), the various of the users 46, 48, 50 can modulate
`their signals into one or more of the available frequency
`bands 42 on a time slot-by-slot 44 basis in order to effect
`optimum scheduling of the users Within the medium 40 to
`ef?ciently make use of the available time-frequency medium
`40. The actual positioning (scheduling) of the various speed
`users 46, 48, 50 Within the overall medium may be deter
`mined based on such factors as individual user demand, the
`relative numbers of loW speed/medium speed/high speed
`users, and the like.
`One Way to effect the slicing of the transmission medium
`40 and to implement positioning of the users 46, 48, 50
`Within the medium is to provide a central control 100 to
`maintain or otherWise keep a lookup table containing the
`status of the availability of space Within the medium 40
`according to frequency band allocations 42 and time slots
`44. The central control 100 may then aWard particular
`time-frequency slice 52 allocations to the individual users
`46, 48, 50 based on such factors as the amount of the
`medium 40 requested by the users and/or the amount of
`medium 40 already allocated to users. Individual users may
`thus align themselves Within their assigned time-frequency
`slices 52 through appropriate signal con?guration and/or
`modulation. Based on the availability of the medium 40,
`central control 100 can thus allocate particular time
`frequency slices 52 to a given user so as to anticipate
`“future” requests Which Will be made by users 46, 48, 50 so
`
`15
`
`25
`
`35
`
`45
`
`55
`
`

`
`6,018,528
`
`5
`as to best optimize full use of the overall medium 40. The
`control 100 can anticipate such requirements, for instance,
`through use of probabalistic studies, historical or projected
`load requirements, and the like, as normally maintained by
`individual service providers. Another Way to effect use
`spectrum of the medium 40 is through random assignments
`of users 46, 48, 50 to the available time-frequency slices 52.
`Other Ways of effecting slicing and scheduling in accordance
`With the system and method of the invention can be readily
`envisioned or otherWise arrived at by those skilled in the art.
`As Will be appreciated, through scheduling, the time
`frequency spectrum 40 can be ?lled in a more ef?cient
`manner than possible With the Universal-Time-Slot
`approach. Unlike a pure TDMA approach, a common band
`Width is not required, so that the system and method can
`schedule cost-ef?cient entry points for loWer speed users 50.
`That is, unlike TDMA, users are capable of operating at their
`oWn access rates While still being able to share the overall
`time-frequency domain 40 With users operating at different
`access rates. As shoWn in FIG. 5, several loW-speed users 50
`can be scheduled to transmit on different frequencies 42 in
`the same designated time slot 44. For instance, loW speed
`users S, J and T occupy the same time slot S6. During certain
`other time slots 44, then, a smaller number of high-speed
`users 46 may be scheduled to transmit.
`Moreover, unlike a pure FDMA approach, a given band
`Width 42 can be occupied by multiple users (for instance,
`users G, B, H, P, S for band F6). Thus, the system and
`method provide a large degree of ?exibility in ef?ciently
`packing the time-frequency spectrum 40 and making use of
`the entire domains.
`Oftentimes, it is advantageous that high-speed users 46 be
`assigned contiguous frequencies 42. Such contiguous
`assignments eliminate the need for guard bands betWeen the
`frequencies assigned to a given user. Depending on the
`modulation scheme, hoWever, certain adjacency require
`ments may be relaXed. For instance, as Will be appreciated,
`users modulating their signals according to a “multi-tone”
`scheme may not require contiguous frequency assignments
`in order to transmit their data. As those skilled in the art Will
`discern, tones represent multi-bit symbols, With each tone
`toggling at a rate corresponding to the bandWidth of one
`frequency band. Thus, With multi-tone transmission tWo bits
`can be transmitted as one 4-ary symbol using 2-tone modu
`lation instead of tWo symbols on a binary channel.
`FIG. 6 thus depicts a variation of the time-frequency
`slicing method of the invention Where noncontiguous fre
`quency arrangements may be employed. For instance,
`higher-speed users 46 operating on multi-tone modulation
`may bene?t from non-contiguous frequency arrangements.
`Here, a particular high speed user B (designated on FIG. 6
`by numeral 54) has been assigned tWo non-contiguous
`frequency assignments (“slices”) F0 and F5—F6 in the
`bandWidth, rather than the single contiguous assignment
`F4—F6 that the same user B might have employed Without
`multi-tone modulation as depicted in FIG. 5. Each of the
`respective tones modulated by the user (here, B) can occupy
`a respective frequency assignment Without the necessity for
`contiguous assignments.
`An eXample of a multi-tone approach includes current
`channeliZed cellular systems, for instance, cellular telephone
`systems, cellular data systems, or the like, to provide higher
`bandWidth to some users. The higher bandWidth is accom
`plished by allocating multiple channels to each higher-speed
`user. Since the allocations do not need to be contiguous,
`more users can perhaps be accommodated than With con
`
`10
`
`15
`
`25
`
`35
`
`45
`
`55
`
`65
`
`6
`tiguous assignments (FIG. 5). The blocking probability may
`be reduced compared to the contiguous assignments of the
`time-frequency approach as described in FIG. 5.
`Thus, it Will be realiZed that higher data rates Will be
`available to higher-speed users 46 by signaling on a com
`bination of tones, Whereas the loWer-speed user 50 Would
`occupy only a single frequency slice of the bandWidth. The
`transmission speed of a user can thus determine the number
`of tones and, thus, the number of frequencies 42 allocated
`for that user. These tones may be scheduled in possibly
`non-contiguous frequency slots Within one or more time
`slots, as, for eXample, for user B in FIG. 6. In fact, it has
`been found that spreading the frequency allocations of a
`high-speed user may offer some propagation bene?ts (e.g.,
`a reduction in the degradation from frequency-selective
`multipath fading).
`It Will be understood, of course, that the single
`transmitter-receiver arrangement as utiliZed in FIG. 5 Will
`not be employed by high-speed users in multi-tone trans
`mission in order to obtain this scheduling advantage. Here,
`higher speed users may need to employ multiple
`transmitters, one for each frequency slice that has been
`assigned to that particular user. HoWever, it Will be under
`stood that as opposed to contiguous transmissions entailing
`the entire frequency spectrum, for non-contiguous multi
`tone transmissions, the base station receiver itself may be
`simpli?ed, in that only a ?Xed number (“n”) tones in speci?c
`frequency bands 42 Will need to be received, so that only a
`single, loW bit rate transmitter/receiver pairing may need to
`be used. It Will also be realiZed that the m-ary components
`may be modulated by a spectrally ef?cient scheme or by a
`constant envelope scheme such as constant poWer PSK.
`Higher-level modulations are also possible in the system and
`method according to the invention.
`Other applications of the scheduling method and system
`according to the invention are also possible. As Will be
`appreciated to those skilled in the art, in addition to the
`TDMA and FDMA multiple access architectures, a “Code
`Division Multiple Access” (CDMA) system may also be
`employed in an effort to permit multiple access to the
`communications transmission medium. Abrief revieW of the
`principles of CDMA architecture Will serve to better appre
`ciate the applicability of the principles of the system and
`method according to the invention to that architecture.
`In CDMA, individualiZed transmissions are not strictly
`separated by frequency (as in FDMA) or strictly separated
`by time (as in TDMA). Rather, transmissions in CDMA are
`permitted to controllably interfere With one another by
`sharing the same frequency spectrum at the same time. By
`assigning a special, unique code to each of the separate
`transmissions occupying the CDMA medium, each particu
`lar transmitter-receiver pair (Which operates according to a
`respective code) may decode the appropriate transmission
`occupying the common channel from among the other
`signals occupying that same channel.
`One Way to implement CDMA is via “Direct Sequence
`Spread Spectrum”, in Which users are assigned codes of
`small cross-correlation. For eXample, this code set, large but
`?nite, may be composed of different phases of a long
`PN-sequence. When users access the channel, they multiply
`their modulated data stream by their assigned code. The
`code rate, Which is considerably higher than the data bit-rate,
`is referred to as the chip-rate. At the receiving end, the
`destination multiplies the received signal by a replica of the
`source code to recover the original signal.
`As those skilled in the art Will realiZe, CDMA support for
`multiple access stems from the fact that the cross-correlation
`
`

`
`10
`
`15
`
`25
`
`35
`
`7
`between tWo different codes is small. Thus, if a signal
`encoded at one code (C1) is decoded With a different code
`(C2), the result appears to the receiver as noise. The limi
`tation of the scheme (i.e., the maximum number of users that
`can utilize the multiple access channel) depends on the total
`amount of “noise” contributed by “interfering” users to the
`detected signal. In other Words, the more users simulta
`neously transmitting on the channel, the greater the level of
`interference that Will exist Within the medium. The Signal
`to-Interference ratio (S/I) determines the Bit-Error-Rate
`(BER) performance of the system.
`In the spectral domain, the multiplication of the data by
`the fast bit-rate code corresponds to spreading the data
`spectral components over a broader spectrum. Thus, a larger
`spectrum is required to convey the transmission. HoWever,
`because of the multiple-access feature, a number of users
`may co-exist at any time on the channel. The ratio of the
`unspread and the spread signals is called the processing gain,
`GP, and GP=2RC/Rb Where RC and Rb are the chip and the
`data bit-rates, respectively. The larger the processing gain,
`the less “noise” contribution any user has on the other users’
`signals.
`The principles underlying the system and method of the
`invention Will serve to enhance usage of the CDMA
`medium. The resource space might be sliced into a “time
`code” space, a “frequency-code” space or, if vieWed in three
`dimensions, into a “time-frequency-code” space. Thus, it
`Will be appreciated that the scheduling approach according
`to the system and method of the invention can also be used
`in the CDMA domain to improve resource usage.
`FIG. 7 depicts application of a “time-code” slicing
`method as applied to transmissions in the CDMA domain.
`FIG. 8 depicts a “frequency-code” slicing approach. As
`before, a plurality of different speed users 46, 48, 50 are
`contemplated. The overall medium 40‘ is partitioned into a
`plurality of individual, discrete “codes” (43) either over the
`time (44) domain (FIG. 7) or frequency band 42 domain
`(FIG. 8), accounting for the relative use of the available code
`space Which is contained Within the overall medium 40‘.
`The term “code space” is used to denote the overall set of
`all possible codes for assignment to user transmission
`employing, for instance, a “family” of codes acceptable for
`purposes of cross-correlation. Auser requiring a large degree
`of code space—for instance, users G, B, M, Q, F—can be
`granted code space in at least tWo Ways. For purposes of
`illustration and not of limitation, examples of possible code
`space allocations are presented in FIGS. 7, 8 and 9. In FIGS.
`7 and 8, users B and G, for instance, require a relatively large
`quantity of code space and as such are granted a plurality of
`individual codes 43 across time slots (FIG. 7) or frequency
`bands (FIG. 8). The plurality of individual codes are col
`lectively representative of a larger quantity of code space
`contained Within the overall medium 40‘.
`An alternative approach is illustrated in FIG. 9. Here, a
`user may be allocated codes of differing length 120. The
`relative length of a given code is inversely related to the
`quantity of code space to be occupied by a given user. For
`instance, in FIG. 9, user Ais assigned a longer code C5 than
`user G (code C6). As illustrated in FIG. 9, the relative
`“height” of the code space occupied by those users is
`indicative of the quantity of code space occupied by them;
`here, user A, Who has been assigned a longer code (C5) than
`user G (code C6) occupies less code space than user G. In
`this manner, optimum use of the overall code space embod
`ied Within the medium 40‘ can be achieved.
`It can be seen in FIG. 9 that the system and method
`provide for ef?cient reuse of the available codes based on the
`
`45
`
`55
`
`65
`
`6,018,528
`
`8
`temporal occupancy requirements of a given user Within the
`medium. For example, it can be seen that code C3 can be
`reused a number of times—here, by users C, J, N,
`R—because each of those users do not occupy any common
`portion of the overall code space located in the medium 40‘
`at the same time.
`In general, it can be