`
`SONY Exhibit 1008 - 0001
`
`
`
`FEBRUARY 2000 VOL. 7 NO. I
`
`EEPers0nal
`ommntcattons
`
`network of
`n'chmneus- 4
`peer den-"ices
`
`Guest Editorial:
`
`.
`
`.
`
`Connectivity and Application Enablers for
`Ubiquitous Computing and Communications
`Chatschik Bisdikian, Jaap C. Haartsen, and Parviz Kermani
`
`8
`Remembering Mark Weiser: Chief Technologist‘, Xerox PARC
`Roy Want
`
`1 1
`
`IrD/4: Past Present and Future
`Stuart Williams
`20
`HomeRF: Wireless Networking for the Connected Home
`Kevin J. Negus, Adrian P. Stephens, and Jim Lansford
`
`28
`
`The Bluetooth Radio System
`Jaap C. Haartsen
`
`37
`
`Paving the Way for Personal Area Network Standards:
`An Overview ofthe IEEE P802.I5 Working Group for
`Wireless Personal Area Networks
`
`Thomas M. Siep, Ian C. Gilford, Richard C. Braiey, and Robert F. Heile
`
`44
`
`System Support for Mobile, /ldaptive /Ipplications
`Brian Noble
`
`f;,:‘j;:l!““m‘l”"’ The Slack
`
`Editor’s Note — 4
`Index to Articles: I999 — 50
`
`IEEE. Personal Con1munjcation.s ' February 2000
`
`SONY Exhibit 1008 - 0002
`
`
`
`EDITOR’S NOTE
`
`A l is my utmost honor and privilege
`to serve our IEEE Personal’ Contnmnicatians
`
`community starting with this issue of our
`magazine. At first, on behalf of our Editorial
`Board and the IEEE Communications Soci-
`
`ety, Iwould like to extend my sincere grati~
`tude to Tom La Porta, who served as
`Editor-in-Chief from 1997 to 1999. Tom will
`
`tion of the possibilities and potentials wire-
`less and mobile technologies will bring to us.
`It is expected that by 2003 the number of
`cellular subscribers will be equal to the
`number of wired phone subscribers. Obvi-
`ously, new evolving architectures and stan-
`dards, such as GPRS and EDGE, and
`third-generation systems will play a major
`role in this transition, enabling a high-speed,
`ubiquitous wireless Internet. After all, it will
`be a totally new world when one billion peo-
`pie carry a personal gateway to the Internet
`with integrated voice and data support.
`Today, wireless technologies have crossed
`the cost, integration, and power consump-
`tion barriers, and are a key factor in defining new directions
`not only it: enterprise but also in consumer markets. Sec-
`or1d~generation cellular technologies are leading this front,
`
`continue to help and guide us as at senior
`advisor or’ the magazine. It will be hard to
`match the high level of leadership quality he
`has established, and I count on him. our
`founding Editor-in-Chief," Hamid Ahmadi,
`our Advisory Board, technical editors, and
`the LEEE editorial team to help me carry my
`responsibility as the new Editor-in-Chief.
`In my view, this is the most exiting time in our area;
`we are starting a new millennium with tremendous anticipa-
`
`MAHMOUD
`NAGHSHINEH
`
`Mark J. Karol. Lueent Te
`ologtes, USA
`Director of Magazine;
`lEditor—in~Cltief
`Mahmoud Naghshineh. [BM Research, USA
`Senior Advise rs
`Hamid Ahmatli. AT&T Labs. USA
`Thomas F. La Porta, Lueer1tT'ecl'tnologics, USA.
`Advisory Board
`Donald Cox. Stanford University. USA
`David Goodman. Rutgers University. USA
`Jorma Lilleberg, Noltia. Finland
`Kaveh Fahlavan, Worcester Polyteeh nit:
`institute. USA
`Mnhndev S-atyanarayanan, CMU, USA
`IEEE Vehicular Technology Liaison
`‘Theodore Rappaport, Virginia Tech, USA
`
`Mike Liu, 0 in State University. USA
`IEEE Computer Society Liaison
`Technical Editors
`Urnesh Amin, AT&T Wireless Services. USA
`B. R. Badrinalh. Rutgers University. USA
`Pravin Bha
`at. IBM Research. USA
`Kwan —C.‘heng hct1,'l'sing Hue Univ., Taiwan
`Si Tak tanley Chin, Air‘i'ouch lntemntional. UK
`Andrea Goldsmith. Caltech. USA
`Paul Gou h. Philips Research. ‘UK
`Davide Grillo, ondazionc Ugo Bordoni, Italy
`.l:titp Httartsen, Ericsson, Sweden
`Tnkeshi Hattori. NTT. Japan
`Ravi Jain. B-ellcore, USA
`Jose h Kahn. UC Berlteley, USA
`Parviz
`ermani. IBM Research, USA
`Richard LaMaire. EBM Research, USA
`Murray Mnzer, Open Group Research Inst. USA
`Sergio Paiazzo, Universtty of Cattmia, Italy
`Ramachandran Ramos, Lucent Technologies,
`Bell
`bs, USA
`Bill-Schilit, FX Palo Alto Lab, lnc.. USA
`Thomas Y. (L Woo, Lucent Technolo ‘es. USA
`Yaeov Yacobi, Microsoft Corp., 65.’.
`Michele Zora’. University of CA San Diego. USA
`Department Editors
`_ Book Rerietvs
`Snshadrt Mohan, Bellcorc, USA
`Cortfvrertce Review
`'|'horna.s Y. C. Woo, Lunent Technologies, USA
`Smmtirt
`rlte Uterrttt-‘J’:
`_
`YI-Bing Lin. Nation Chiao Tung Univ., Taiwan
`IEEE Production Staff
`Joseph Milizzn, Manager. Print &
`Electronic Publishing
`Catharina: Kemelmacher. Production Editor
`Eric Levine. Advertising Sales Manager
`Joanne O'Rourke. Staff Assistant
`Susan Lari e, La out Editor
`Jennifer Porcel o. E itorial Assistant
`
`IEEEPer:sonaI
`Communications
`1'»: MM:.Arm: on w-nn.:n cauuumcomnm up NI.‘lN."JMk<Nl'.
`
`2000 Communications Society
`Board of Governors
`
`Officers
`1. Roberto De lv1arca.l‘mn'rlr-rt:
`Thomas i. Plcvyak. Pas: I’:-esidertr
`Curtis A. Siller. VP-il'echrticol/lcrivitiex
`Horst Bessai, VP—Membet:nlu'p Services
`Douglas N. Zuckcrman, VP-
`Mcrniieisitip Darelopnrmr
`Federico Tosco. I/P—S‘uct'ety Relations
`Harvey Freeman, Tl"t’fl£lll'£r
`John M. Howell, St'r'rL-r.trr_\-
`
`Mt.-rnbers«at-Large
`Ctai-myrzaaa
`Gerhard Fcttweis
`Paul Harlmann
`Michael Kincaid
`William R. Robinson
`Class of 200!
`Laura Cerchio
`Leonard Cimini
`Rohcrta Cohen
`Willinnt Tranter
`Class of2t.'.‘l02
`Tomnnori Aoynma
`Ales Gelrnatn
`Roclt Guerin
`Byeong Lee
`2000 IEEE Officers
`Bruce A. Eisenstein, Pmridertt
`Joel H. Snyder. President-Elect
`David J. Kemp. .5'cr-rr:mt;r
`David A. Conner. '."'nmsrrrt-.'r
`Kenneth R. Lnker. Pro"! Presidertr
`Daniel J. Sencse, .Ext.'r:IJ.tr'w.'Direclar
`Tom Rowbotham,Dtmctor. Diuiuioix ill
`
`llllllllll
`
`IEEE Personal Communications - The Maga-
`zine of Wireless Communications and
`Neltvorking {ISSN ll'I'.r'n-9'-Jlfi] is published
`himonthlyhy The Institute of Elcctricttl and
`Electronics l£t:git1ncrs.lnc. I-lcadqu.-srtcrs add rest:
`IEEE. 3- Park Avenue. |'.lth Floor. New York.
`NY liltllt--5997: tel: 2l2—?tlS-8900; l':u:: 2I2-?05-
`8999: e-mail: c.ItemsImucherCtiJcomsnc.org.
`Responsibility for the contents rests upon authors
`ofsigncd articles and not the IEEE or its numbers.
`Unless otherwise specified. the IEEE neither
`endorses nor sanctions any positions or actions
`espoused in JEEE Pcrrottai Coulnluttirrrtintis.
`Annual so hscription: Men-ibersubscriptinnzszs
`peryc ar: Nonmember subscription prices available
`on request. Single copy: Sill for members and $30
`for nonmembers.
`
`Editorial correspondence: Manuscripts for
`consitlerztlion may he strhmitled to the Etlll0r-iI‘|-
`Chief: Thomas F. La Port:I. Ltlccnl Technolo-
`gies. I:1c., 101 Crawford Corners Road.Rm -‘IF-Sl 9.
`Holn1del.N.l ilTl'33. Electronic suhntissionstnrtyhe
`Sent in postscript to: tlp@be|l-l:tbs.con1.
`Copyright and reprint permissions: Abstract-
`ing is pt! nrtitted with credit to the source. Lib rarios
`are permitted to photocopy beyond the limits
`of US. Copyright law [or private use of patrons:
`those post-197? articles that carry a code on the
`bottom of the first page provided the percopy fee
`indicated in the code ispnid through the Copyright
`Clearance Center. 222 Rosewood Drive, Dan-
`vets. MA I] I 923. For other copying. reprint, or
`republication permission. write to Director,
`Publishing Services. at IEEE I-lcadquartt-rs. All
`tigltls reserved. Copyright Q Ztltlli by The Insti-
`tute DI‘ Elect rical and Eiectmnics Engineers, Inc,
`Postmaster: Send address changes to {BEE
`Pcrsorr.-J!‘ Comnrttrticnriorls. IEEE. 445 Hoes
`Lane. Piscatuwaly. NJ 08855-1331; or email to
`:1tldress.L‘hangc(_rI'iece.org. Printed in USA. Peri-
`otliculs postage paid at New York, NY and at addi-
`tional mailing ol'fices.Cant:tdian GST#125634l88_
`Subscriptions: 5:: ml orders, address changes to:
`1EEE Service Center. 445 Hoes Lane, Fiscal-
`nway. NI (18855-1331‘. tel.: 908381-GU51}.
`Advertising: Advertising is accepted at the dis-
`cretion of the publisher. Address correspondence
`In: Advertising Manager. IEEE Per.rom1t‘Com-
`rrrtrtticnriottr. 305 East 47th Street. New York,
`NY llllll?-Z394.
`
`JEEE Personal Communications - February 2000
`
`SONY Exhibit 1008 - 0003
`
`
`
`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`Abstract
`A few years ago it was recognized that the vision of a truly low-cost. low-power radio-based cable replacement was feasible. Such a ubiquitous
`link would provide the basis for portable devices to communicate together in an ad hoc fashion by creating personal area‘ networks which have
`similar advantages to their office environment counterpart, the local area network. Bluetoothm is an effort by a consortium of companies to‘
`design a royalty-free technology specification enabling this vision. This article describes the radio system behind the Bluetooth concept: Designing
`an ad hoc radio system for worldwide usage poses several cltallcnges. The article describes the critical system characteristics and motivates the
`design choices that have been made.
`
`The Bluetooth Radio System
`
`JAAP C. HAARTSEN, ERICSSON RADIO SYSTEMS B.V.
`
`n the last decades, progress in
`microelectronics and very large scale integration (VLSIJ tech-
`nology has fostered the widespread use of computing and
`communication devices for commercial usage. The success of
`consumer products like PCs, laptops, personal digital assis-
`tants (PDAs). cell phones, cordless phones. and their periph-
`erals has been based on continuous cost and size reduction.
`Information transfer between these devices has been cumber-
`some, mainly relying on cables. Recently, a new universal
`radio interface has been developed enabling electronic devices
`to communicate wirelessly via short-range ad hoc radio con-
`nections. The Bluetooth technology — which has gained the
`support of leading manufacturers like Ericsson, Nokia, IBM,
`Toshiba, Intel. and many others — eliminates the need for
`wires, cables, and the corresponding connectors between cord-
`less or mobile phones. modems, headsets, PDAS, computers,
`printers, projectors, and so on, and paves the way for new and
`completely different devices and applications. The technology
`enables the design of low-power, small-sized, low-cost radios
`that can be embedded in existing (portable) devices. Eventual-
`ly, these embedded radios will lead toward ubiquitous connec-
`tivity and truly connect everything to everything. Radio
`technology will allow this connectivity to occur without any
`explicit user interaction.
`This article describes the basic design and technology
`trade-offs which have led to the Bluetooth radio system. We
`describe some fundamental issues regarding ad hoc radio sys-
`tems. We give an overview of the Bluetooth system itself with
`the emphasis on the radio architecture. It explains how- the
`system has been optimized to support ad hoc connectivity. We
`also describe the Bluetooth specification effort.
`
`Ad Hoc Radio Connectivity
`The majority of radio systems in commercial use today are
`based on a cellular radio architecture. A mobile network estab-
`lished on a wired backbone infrastructure uses one or more
`base stations placed at strategic positions to provide local cell
`coverage; users apply portable phones, or more generic mobile
`terminals, to access the mobile network; the terminals main-
`tain a connection to the network via a radio link to the base
`stations. There is a strict separation between the base stations
`and the terminals. Once registered to the network, the tenni-
`nals remain locked to the control channels in the network, and
`connections can be established and released according to the
`control channel protocols. Channel access, channel allocation,
`traffic control, and interference minimization are neatly cott-
`
`trolled by the base stations. Examples of these conventional
`radio systems are the public cellular phone systems like Glob-
`a] System for Mobile Communications (GSM), D-AMPS, and
`IS-95 [1—3], but also private systems like wireless local area
`network (WLAN) systems based on 802.11 or I-IIPERLAN I
`and HIPERLAN 1] [4—6], and cordless systems like Digital
`Enhanced Cordless Telecommunications (DECT) and Person-
`al Hattdyphone System (PHS) [7, S].
`In contrast, in truly ad hoc systems, there is no difference
`between radio units; that is, there are no distinctive base sta-
`tions or terminals. Ad hoc connectivity is based on peer com-
`munications. Therc is no wired infrastructure to support
`connectivity between portable units; there is no central con-
`troller for the units to rely on for making interconnections; nor
`is there support for coordination of communications. In addi-
`tion. there is no intervention of operators. For the scenarios
`envisioned by Bluetooth, it is highly likely that a large number
`of ad hoc connections will coexist in the same area without any
`mutual coordination; that is, tens of ad hoc links must share
`the same medium at the same location in an uncoordinated
`fashion. This is different from ad hoc scenarios considered in
`the past, where ad hoc connectivity focused on providing a sin-
`gle (or very few} network(s) between the units in range [4, 5].
`For the Bluetooth applications, typically many independent
`networks overlap in the same area. This will be indicated as a
`scatter ad hoc environment. Scatter ad hoc environments con-
`sist of multiple networks, each containing only a limited num-
`ber o.f units. The difference between a conventional cellular
`environment, a conventional ad hoc environment, and a scatter
`ad hoc environment is illustrated in Fig. I. The environmental
`characteristics the ad hoc radio system has to operate in have a
`major impact on the following fundamental issues:
`' Applied radio spectrum
`- Determining which units are available to connect to
`- Connection establishment
`- Multiple access scheme
`0 Channel allocation
`° Medium access control
`
`- Service prioritization (i.e., voice before data}
`- (Mutual) interference
`- Power consumption
`Ad hoc radio system have been in use for some time, for
`example, walky-tall-ty systems used by the military, police, fire
`departments, and rescue teams in general. However, the Blue-
`tooth system is the first commercial ad hoc radio system envi-
`sioned to he used on a large scale and widely available to the
`public.
`-
`
`28
`
`1070-99l.fif00f$l0.0fl © 2000 IEEE
`
`IEEE Personal Communications - February 2000
`
`SONY Exhibit 1008 - 0004
`
`
`
`Bluetooth Radio System Architecture
`
`In this section the technical background of the Bluctooth
`radio system is presented. It describes the design tradc—offs
`made in order to optimize the ad hoc functionality and
`addresses the issues listed above.
`
`Radio Spectrum
`The choice of radio spectrum is first determined by the lack
`of operator interaction. The spectrum must be open to the
`public without the need for licenses. Second, the spectrum
`must be available worldwide. The first Bluctooth applications
`are targeted at the traveling businessperson who connects
`his/her portable devices wherever he/she goes. Fortunately,
`there is an uniicensed radio band that is globally available,
`This band, the Industrial, Scientific, Medical (ISM) band, is
`centered around 2.45 G1-I2 and was formerly reserved for
`some professional user groups but has recently been opened
`worldwide for commercial use. In the United States, the band
`ranges from 2400 to 2483.5 MHz, and the FCC Part 15 regu-
`lations apply. In most parts of Europe} the same band is
`available under the ETS-300328 regulations. In Japan, recent-
`ly the band from 2400 to 2500 MHz has been allowed for
`commercial applications and has been harmonized with the
`rest of the world. Summarizing. in most countries of the
`world, free" spectrum is available from 2400 MHZ to 2483.5
`MHz, and harmonization efforts are ongoing to have this
`radio band available truly worldwide.
`The regulations in different parts of the world differ. How-
`ever. their scope is to enable fair access to the radio band by
`an arbitrary user. The regulations generally specify the
`spreading of transmitted signal energy and maximum allow-
`able transmit power. For a system to operate globally, a radio
`concept has to be found that satisfies all regulations simulta-
`neously. The result will therefore be the minimum denomina-
`tor of all the requirements.
`
`interference lmntuntty
`Since the radio band is free to be accessed by any radio trans-
`mitter as long as it satisfies the regulations, interference
`immunity is an important issue. The extent and nature of the
`interference in the 2.45 GHz ISM band cannot be predicted.
`Radio transmitters may range, for example, from 10 dBm
`baby monitors to 30 dBm WLAN access points. With high
`probability. the different systems sharing the same band will
`not be able to communicate. Coordination is therefore not
`possible. More of a problem are the high-power transmitters
`covered by the FCC part 18 rules which include, for example,
`microwave ovens and lighting devices. These devices fall out-
`side the power and spreading regulations of part l5, but still
`coexist in the 2.45 GI-E2 ISM band. In addition to interference
`from external sources, co-user interference must be taken into
`account, which results from other Bluetooth users.
`Interference immunity can be obtained by interference
`suppression or avoidance. Suppression can be obtained by
`coding or direct-sequence spreading. However, the dynamic
`range of the interfering and intended signals in an ad hoc,
`uncoordinated radio environment can be huge. Taking into
`account the distance ratios and power differences of uncoordi-
`nated transrnitters, near-far ratios in excess of 50 dB are no
`exception. With desired user rates on the order of 1 Mb/s and
`beyond, practically attained coding and processing gains are
`inadequate-. Instead, interference avoidance is more attractive
`
`4’ In Frtmce and .S‘ptrin the exact location oftiie band dijferfi. arid The Wild
`is -smaiier.
`
`IEEE Personal Communications ' February 209.0
`
`I Figure I. Topologies for: it} cellular radio systeins with squares
`rcpresentirtg stationary base srariorts; b) conventiorm! ad hoc
`systems,‘ and C) scatter ad hoc ntrteins.
`
`since the desired signal is transmitted at points in frequency
`and/or time where interference is low or absent. Avoidance in
`time can be an alternative if the interference concerns a
`pulsed jarnmer and the desired signal can be interrupted.
`Avoidance in frequency is more practical. Since the 2.45 GHz
`band provides about 80 MHZ of bandwidth and most radio
`systems are band-limited, with high probability a part of the
`radio spectrum can be found where there is no dominant
`interference. Filtering in the frequency domain provides the
`suppression of the interferers at other parts of the radio band.
`The filter suppression can easiiy arrive at 50 dB or more.
`
`Multiple Access Scheme
`The selection of the multiple access scheme for ad hoc radio
`systems is driven by the lack of coordination and the regula-
`tions in the ISM band. Frequency-division multiple access
`(FDMA) is attractive for ad hot: systems since channel orthog-
`onality only relies on the accuracy of the crystal oscillators in
`the radio units. Combined with an adaptive or dynamic chan-
`nel allocation scheme, interference can be avoided. Unfortu-
`nately, pure FDMA does not
`fulfill
`the spreading
`requirements set in the ISM band. T'inJe—divi.siorI multiple
`access (TDMA) requires strict time synchronization for chan-
`nel orthogonality. For multiple collocated ad hoc connections,
`maintaining a common timing reference becomes rather cum-
`bersome. Code-division multiple access (CDMA) offers the
`best properties for ad hoc radio systems since it provides
`spreading and can deal with uncoordinated systems. Direct
`sequence (DS)-CDMA is less attractive because of the near-
`far problem which requires coordinated power control or
`excessive processing gain. In addition, as in TDMA, DS-
`CDMA channel orthogonality requires a common timing ref-
`erence. Finally, for higher user rates, rather high chip- rates
`are required, which is less attractive because of the wide
`bandwidth (interference immunity) and higher current con-
`sumption. Frequency-hopping (FH)—CDMA combines a num-
`ber properties which make it the best choice for ad hoc radio
`systems. On average the signal can be spread over a large Ere»
`quency range, but instantaneously only a small bandwidth is
`occupied, avoiding most of the potential interference in the
`ISM band. The hop carriers are orthogonal, and the interfer-
`ence on adjacent hops can effectively be suppressed by filter-
`
`SONY Exhibit 1008 - 0005
`
`
`
`525 [15
`
`I Figure 2. An illustration ofrhe FH/TDD channel applied in Blttetooth.
`
`ing. The hop sequences will not be orthogonal (coordination
`of hop sequences is not allowed by the FCC rules anyway),
`but narrowband and co-user interference is experienced as
`short interruptions in the communications, which can be over-
`come with measures at higher-layer protocols.
`Biuetooth is based on FH-CDMA. In the 2.45 GHz ISM
`band, a set of 79 hop carriers have been defined at a I MHz
`spacing? The channel is a hopping channel with a nominal
`hop dwell time of 625 us. A large number of pseudo-random
`hopping sequences have been defined. The particular
`sequence is determined by the unit that controls the Fl-l chan-
`nel, which is called the master. The native clock of the master
`unit also defines the phase in the hopping sequence. All other
`participants on the hopping channel are slaves; they use the
`master identity to select the same hopping sequence and add
`time offsets to their respective native clocks to synchronize to
`the frequency hopping. in the time domain, the channel is
`divided into slots. The minimum dwell time of 625 its corre-
`sponds to a single slot. To simplify implementation, full-
`duplex communications is achieved by applying time-division
`duplex (TDD). This means that a unit alternately transmits
`and receives. Separation of transmission and reception in time
`effectively prevents crosstalk between the transmit and receive
`operations in the radio transceiver, which is essential if a one-
`chip implementation is desired. Since transmission and recep-
`tion take place at different time slots, transmission and
`reception also take place at different hop carriers. Figure 2
`illustrates the Fl-IITDD channel applied in Bluetooth. Note
`that multiple ad hoc links will make use of different hopping
`channels with different hopping sequences and may have mis-
`aligned slot timing.
`
`The Modulation Scheme
`In the ISM band, the signal bandwidth of F1-1 systems is limit-
`ed to 1 MHz. For robustness, a binary modulation scheme was
`chosen. With the above-mentioned bandwidth restriction, the
`data rates are limited to about 1 Mbls. For Fl-I systems and
`support for bursty data traffic, a noncoherent detection
`scheme is most appropriate. Bluetooth uses Gaussian-shaped
`frequency shift keying (FSK) modulation with a nominal mod»
`ulation index of It = 0.3. Logical ones are sent as positive fre-
`quency deviations, logical zeroes as negative frequency
`deviations. Demodulation can simply be accomplished by a
`limiting FM discriminator. This modulation scheme allows the
`implementation of low-cost radio units.
`
`Medium Access Control
`Bluetooth has been optimized to allow a large number of
`uncoordinated communications to take place in the same
`area. Unlike other ad hoc solutions where all units in range
`
`-’ Cttrrt-ntiy, filr Frttnce and Spain a reduced set of2'3 hop centers has been
`defined at :1 1 MHz carrier spacing.
`
`share the same channel. Biuetooth has been
`designed to allow a large number of independent
`channels, each channel serving only a limited num-
`ber of participants. With the considered modulation
`scheme, a single FH channel in the ISM band only
`supports a gross bit rate of l Minis. This capacity has
`to be shared by all participants on the channel. The-
`oretically, the spectrum with ‘)9 carriers can support
`79 Mbfs. In the user scenarios targeted by Blue-
`tooth. it is highly unlikely that all units in range
`need to share information among all of them. By
`using a large number of independent 1 Mbfs chan-
`nels to which only the units are connected that real-
`ly want to exchange information, the 80 MHz is exploited
`much more effectively. Due to nonorthogonality of the hop
`sequences, the theoretical capacity of 79 Mbfs cannot be
`reached. but is at least much larger than 1 Mb/s.
`An Fl-I Bluetooth channel is associated with a piconet. As
`mentioned earlier, the piconet channel is defined by the identi-
`ty (providing the hop sequence} and system clock {providing
`the hop phase) of a master unit. All other units participating in
`the piconet are slaves. Each Bluetooth radio unit has -a free-
`running system or native clocl-:. There is not a common timing
`reference, but when a piconet is established, the slaves add off-
`sets to their native clocks to synchronize to the master. These
`offsets are released again when the piconet is cancelled, but can
`be stored for later use. Different channels have different mas-
`ters and therefore also different hopping sequences and phases.
`The number of units that can participate on a common channel
`is deliberately limited to eight {one master and seven slaves) in
`order to keep a high-capacity link between all the units. It also
`limits the overhead required for addressing. Bluetooth is based
`on peer communications. The master./slave role is only attribut-
`ed to a unit for the duration of the piconct. When the piconet
`is cancelled. the master and slave roles are cancelled. Each unit
`can become a master or slave. By definition, the unit that estab-
`lishes the piconet becomes the master.
`In addition to defining the piconet, the master also controls
`the traffic on the piconet and takes care of access control.
`Access is completely contention free. The short dwell time of
`625 us only allows the transmission of a single packet. A con-
`tention-based access scheme would provide too much over-
`head and is not efficient in the short dwell time Bluetooth
`applies. In Bluetooth, the master implements centralized con-
`trol; only communication between the master and one or more
`slaves is possible. The time slots are alternately used for mas-
`ter transmission and slave transmission. In the master trans-
`mission, the master includes a slave address of the unit for
`which the information is intended. In order to prevent colli-
`sions on the channel due to multiple slave transmissions, the
`master appliesa polling technique: for each slave-to-master
`slot, the master decides which slave is allowed to transmit. This
`decision is performed on a per—slot basis: only the slave
`addressed in the master-lo- slave slot directly preceding the
`slave-to-master slot is allowed to transmit in this slave-to-may
`ter slot. If the master has information to send to a specific
`slave, this slave is pulled implicitly and can return information.
`If the master has no information to send, it has to poll the
`slave explicitly with a short pol] packet. Since the master
`schedules the traffic in both the uplink and downlink. intelli-
`gent scheduling algorithms have to be used that take into
`account the slave characteristics. The master control effectively
`prevents collisions between the participants on the piconet
`channel. Independent collocated piconets may interfere when
`they occasionally use the same hop carrier. A type of ALOHA
`is applied: information is transmitted without checking for a
`clear carrier (no listen-before-talk). If the information is
`
`IEEE Personal Communications - February 2000
`
`SONY Exhibit 1008 - 0006
`
`
`
`received incorrectly, it is retransmitted at the next
`transmission opportunity {for data only]. Due to the
`short dwell time, collision avoidance schemes are
`less appropriate for FH radio. For each hop, differ-
`ent contenders are encountered. Backoft'mecha-
`nisms are therefore less efficient.
`
`72 bits
`
`_ 54 bits
`
`0-2745 bits
`
`gum
`
`code
`
`Packet
`header
`
`Payload
`
`i Figure 3. The format ofpackets appiieol in Blrretoolh.
`
`Packet-Based Communications
`The Bluetooth system uses packet—based transmission: the
`information stream is fragmented into packets. In each slot,
`only a single packet can be sent. All packets have the same
`format, starting with an access code, followed by a packet
`header, and ending with the user payload (Fig. 3}.
`The access code has pseudo-random properties and is used
`as a direct-sequence code in certain access operations. The
`access code includes the identity of the piconet master. All pack-
`ets exchanged on the channel are identified by this master iden-
`tity. Only if the access code matches the access code
`oorresponcling to the piconet master will the packet be accepted
`by the recipient. This prevents packets sent in one piconet false-
`ly being accepted by units of another piconet that happens to
`land on the same hop carrier. In the receiver, the access code is
`matched against the anticipated code in a sliding correlator.
`This correlator provides the direct-sequence processing gain.
`The packet header contains link control information: a 3-bit
`slave address to separate the slaves on the piconet, a 1-bit
`acknowledgment/negative acimowledgrncnnt (ACKtNACK} for
`the automatic repeat request (ARO) scheme, a 4-bit packet type
`code to define 15 different payload types, and an 8-bit header
`error check (HEC) code which is a cyclic redundancy check
`(CRC) code to detect errors in the header. The packet header is
`limited to 18 infonnation bits in order to restrict the overhead.
`The header is further protected by 113 rate forward error correc-
`tion (FEC) coding. Blue-tooth defines four control packets:
`' The ID or identification packet: Only consists of the access
`code; used for signaling
`- The NULL packet: Only has an access code and a packet
`header; used if link control information carried by the
`packet header has to be conveyed
`I The POLL packet: Similar to the NULL packet; used by the
`master to force slaves to return a response
`' The FHS packet: An FH-synchronization packet; used to
`exchange real-time ciock and identity information between
`the units; contains all the information to get two units hop
`synchronized
`The remaining 12 type Codes are
`used to define packets for syn-
`chronous and asynchronous services.
`These 12 types are divided into three
`segments. Segment 1 specifies pack-
`ets that fit into a single slot, segment
`2 specifies 3-slot packets, and seg-
`ment 3 specifies 5-slot packets. Multi-
`slot packets are sent on a single-hop
`carrier. The hop carrier which is valid
`in the first slot is used for the remain-
`der of the packet; therefore, there is
`no frequency switch in the middle of
`a packet. After the packet has been
`sent, the hop carrier as specified by
`the current master clock value is used
`(Fig. 4). Note that only an odd num-
`ber of multislot packets have been
`defined, which guarantees that the
`TXIRX timing is maintained.
`On the slotted channel, syn-
`chronous and asynchronous links
`
`525 95'
`
`‘(kl
`
`1-5;
`
`E
`
`I
`
`_
`:
`
`I
`
`.
`
`_
`
`.
`
`I
`
`have been defined, as will be further explained later. The
`interpretation of packet type is different for synchronous and
`asynchronous links. Currently, asynchronous links support
`payloads with or without a 2/3-rate FEC coding scheme. In
`addition, on these links single—slot, three-slot, and five-slot
`packets are available. The maximum user rate that can be
`obtained over the asynchronous link is 723.2 kbfs. In that
`case, a return link of 57.6 kb/S can still be supported. Link
`adaptation can be applied on the asynchronous link by chang-
`ing the packet length and FEC coding depending on link con»
`ditions. The payload length is variable and depends on the
`available user data. However, the maximum length is limited
`by the minimum switching time between RX and TX, which
`is specified at 200 us. This switching time seems large, but
`allows the use of open-loop voltage controlled oscillators
`(VCOs) for direct modulation‘ and provides time for packet
`processing between RX and TX;
`this is also discussed later.
`For synchronous links, only single-slot packets have been
`defined. The payload length is fixed. Payloads with U3-rate
`FEC. 2/3-rate, or no FEC are supported. The synchronous
`[ink supports a full-duplex link with a user rate of 64 kh/s in
`both directions.
`
`Physical Link Definition
`The Bluetooth link supports both synchronous services such
`as voice traffic, and asynchronous services such as bursty data
`traffic. Two physical link types have been defined:
`- The synchronous connection-oriented (SCO} l.i.n.k
`- The asynchronous conncctionless (ACL) link
`The SCO link is a point-to-point link between the master
`and a singie slave. The link is established by reservation of
`duplex siots at regular intervals. The ACL link is a point-to-
`muitipoint link between the master and all the slaves on the
`piconet. The AC]. link can use all of the rernaining slots on
`the channel not used for SCO links. The traffic over the
`ACL link is scheduled by the master. The slotted structure
`of the piconet channel allows effective mixing of the syn-
`