Abstract
`This article provides an overview of the flexible, high-performance packet data channel that has been designed for high-rate
`packet data services over IS-136 TDMA channels. To achieve the highest data rates in the limited 30 kHz channel bandwidth, the
`packet data channel is designed for adaptive modulation and, in addition to a fixed coding mode, permits operation using an
`incremental redundancy mode.
`
`GPRS- 736: High-Rate
`Packet Data Service for North American
`TDMA Digital Cellular Systems
`
`KRISHNA BALACHANDRAN, RICHARD EJZAK, SANJIV NANDA,
`STANISLAV VITEBSKIY, A N D S H I V SETH
`LUCENT TECHNOLOGIES
`
`8-L
`
`urrent North American time-
`division multiple access (TDMA) systems support voice ser-
`vices and circuit data services at a rate limited to 9.6 kb/s.
`High-rate packet data services are desirable for short bursty
`transactions as well as applications such as World Wide
`Web/Internet access, electronic mail, and file transfer. This
`article provides a general overview of the proposed system,
`and describes flexible, high-performance medium access con-
`trol (MAC) and radio resource management procedures that
`have been adopted by the Telecommunications Industry Asso-
`ciation (TIA) for high-rate packet data services over IS-136
`TDMA channels. The 30 kHz spectrum usage, symbol rate,
`and TDMA format (6 time slots every 40 ms) are maintained
`as in the IS-136 standard [ l ] for compatibility with existing
`mobiles and in order to minimize impact on existing infra-
`structure. Assuming these constraints, a new TDMA packet
`data channel (PDCH) is defined to carry both user data and
`information from other higher-layer control and management
`entities.
`Current cellular TDMA systems are designed to achieve
`good coverage over most of a typical cell, and as a result, the
`signal to interference plus noise ratio is sufficient to support
`higher data rates through the use of 8- and 16-level modula-
`tions over a large portion of a typical cell. However, current
`systems rely on the use of only nl4-differential quadrature
`phase shift keying (DQPSK) modulation to achieve modest
`data rates over all operating conditions. The packet data
`MAC and physical layers are designed to increase throughput
`over a significant fraction of the cell area by using coherent 8-
`PSK in addition to d4-DQPSK. The standard is designed to
`permit dynamic adaptation of the modulation scheme based
`on measured carrier-to-interference ratio (C/I). The modula-
`tion schemes are chosen to be the same as those adopted for
`voice services in order to simplify the development of dual-
`mode (voice and packet data capable) mobile stations. Hooks
`are provided in the standard to support a 16-level modulation
`such as 16-quadrature amplitude modulation (QAM), 16-PSK,
`or 16-DPSK in the future.
`
`Three wireless data infrastructure options were investigat-
`ed: Cellular Digital Packet Data (CDPD) infrastructure that is
`widely deployed in the United States; Global System for
`Mobile Communications (GSM) General Packet Radio Ser-
`vice (GPRS) infrastructure that is expected to be deployed
`throughout the world in 1999 and 2000; and the third option
`considered, an Internet service provider (1SP)-like model with
`Point-to-point Protocol (PPP) tunnels to the Internet or cor-
`porate intranets. In order to achieve economies of scale and
`simplify evolution to third-generation systems, the upper lay-
`ers (layer 3 and above) of the packet data protocol stack were
`chosen to be the same as that used by GSM GPRS. GPRS-
`136 is thus a TDMA packet data standard based on GPRS,
`but utilizing 30 kHz for the physical layer and allowing con-
`nection to the American National Standards Institute (ANSI)-
`41 network [2]. GPRS-136 utilizes most of the existing network
`elements from the GPRS network reference model. It adds a
`gateway-mobile switching center (MSC)/visitor location regis-
`ter (VLR) function, which allows connection of the GPRS
`packet data network to the ANSI-41 based mobile circuit-
`switched network. The relevant GPRS specification docu-
`ments can be found in [3-81.
`Another benefit of this choice is that the existing GPRS
`standard is used as a baseline, allowing for quick develop-
`ment. This also enables evolution to the Enhanced General
`Packet Radio Service (EGPRS), which has been proposed to
`the International Telecommunication Union (ITU) as a
`third-generation radio transmission technology for Interna-
`tional Mobile Telecommunications in the year 2000 (IMT-
`2000). The use of EGPRS channels for TDMA packet data
`will be standardized during 1999 and is to be called GPRS-
`136HS.
`
`Organization of the Article
`The following section provides an overview of the service and
`network reference model. The GPRS-based protocol stack is
`described briefly. Provision of voicetdata integration and oper-
`ation with half-duplex and full-duplex terminals is discussed.
`
`34
`
`1070-9916/99/$10.00 0 1999 IEEE
`
`IEEE Personal Communications June 1999
`
`IPR2018-1556
`HTC EX1016, Page 1
`
`

`

`are significantly different from GPRS because of the funda-
`mental differences in the IS-136 and GPRS physical layers.
`The radio resource (RR) sublayer provides data transfer
`and control services to LLC and GPRS-136 mobility manage-
`ment. It consists of three cooperating entities: the MAC enti-
`ty, radio resource management entity (RRME), and broadcast
`management entity (BME).
`The MAC entity provides services directly to LLC through
`a SAP. It supports two multiplexed logical links of different
`priorities, one for normal data transfer and the other for pro-
`viding expedited data delivery services to higher-layer and
`management entities.
`The BME and RRME provide services to the GPRS mobil-
`ity management layer through two other SAPS, and carry out
`both voice and packet-data-related signaling functions. These
`entities use the services of the MAC entity to communicate
`with their peers, and to control the MAC entity. The core
`MAC and RRME procedures are generic, and the interfaces
`can easily be modified to support other protocol stacks.
`Joint VoicelData Operation
`In addition to packet data service, GPRS-136 allows sub-
`scribers to obtain both IS-136 and ANSI-41-based services
`(e.g., circuit-switched voice services, short messaging services,
`intelligent roaming), provided mobile stations support both
`circuit and packet modes. Two new classes of mobile stations
`are supported by GPRS-136: dual mode (i.e., circuit and pack-
`et mode capable) and packet mode only.
`Integration of the GPRS packet network with the ANSI-
`41 circuit-switched network provides a unique challenge.
`GPRS provides its own set of registration, authentication,
`authorization, and mobility management functions which are
`thoroughly integrated with other GSM services such as cir-
`cuit-switched voice service. These functions are implemented
`very differently in the ANSI-41 network. To avoid the com-
`plexity of fully integrating these two sets of functions, GPRS-
`136 provides a method of “tunneling” ANSI-41 signaling
`messages between a mobile terminal and the gateway
`MSCIVLR via the SGSN. These messages are tunneled as
`
`The article next describes the logical and physical structure
`of the PDCH. The allocation of packet paging and packet
`broadcast channels, and PDCH reselection procedures are
`described. Multiple-time-slot operation is also discussed. We
`then describe the MAC layer and uplink media access control
`procedures, including the use of the packet channel feedback
`(PCF) mechanism.
`The article describes the procedures associated with adap-
`tive modulation (DQPSK and coherent 8-PSK), followed by
`the incremental redundancy and fixed coding mode radio link
`protocols (RLPs). The incremental redundancy mode pro-
`vides approximately 15 percent higher throughput at the cost
`of additional receiver memory. We provide detailed time slot
`formats accommodating the fields required by the RLP,
`MAC, and physical layers. The last section provides a brief
`summary.
`
`GPRS- 136 Network
`Reference Model and Operation
`The Packet Data Network Reference Model
`The network elements are as follows:
`Terminal equipment (TE), which typically interfaces with
`the user and contains packet data applications.
`Mobile termination (MT), which interfaces to the TE, and
`terminates the radio interface.
`Base station (BS), which constitutes the interface between
`the network and mobile station, and transfers packet data
`and signaling messages between serving GPRS support
`nodes (SGSNs) and mobile stations in its coverage area.
`SGSN, a packet data switch that routes data packets to
`appropriate mobile stations within its service area.
`Gateway GPRS support node (GGSN), which acts as the
`logical interface between the GPRS-136 network and exter-
`nal packet data networks. It tunnels IP packets from exter-
`nal networks to the SGSN using the GPRS Tunneling
`Protocol (GTP).
`GPRS home location register
`(HLR), accessible from the SGSN
`and GGSN, contains GPRS-136
`subscription and routing informa-
`tion.
`ANSI-41 HLR, accessible from
`the
`serving
`and gateway
`MSCIVLR functions, contains
`subscription and routing informa-
`tion for circuit-switched service.
`ANSI-41 gateway MSCNLR, pro-
`vides functions such as circuit call
`routing and circuit service related
`paging within the GPRS-136 net-
`work.
`ANSI-41 serving MSCIVLR, pro-
`vides circuit switching functions
`for mobile stations in its service
`area.
`Message center (MC), receives
`and accepts requests to deliver
`teleservice messages to the mobile
`subscriber.
`The Protocol Stack (GPRS- 136)
`Figure 2 shows the GPRS-136 pro-
`tocol stack. The MAC and radio
`resource management procedures
`
`w Figure 1. The GPRS-136packzt data network reference model. (Primed interfaces - e.g.,
`Gb’ -indicate ETSI GPRS interfaces that have been modified for GPRS-136.)
`
`IEEE Personal Coinmunications Julie 1999
`
`35
`
`IPR2018-1556
`HTC EX1016, Page 2
`
`

`

`specially marked LLC frames. ANSI-41 registration, authenti-
`cation, authorization, paging, and short message service
`(SMS) messages are delivered transparently (tunneled)
`through the SGSN. This allows the MSC/VLR and SGSN to
`implement and execute these functions independently as nec-
`essary for the proper operation of the ANSI-41 and GPRS
`networks, respectively. The mobile terminal provides the nec-
`essary sequencing of related functions between the ANSI-41
`and GPRS networks.
`The GPRS-136 mobility management layer consists of two
`entities: GPRS mobility management (GMM) and 136 mobili-
`ty management (136MM), which support mobility manage-
`ment functions specific to packet data and circuit-switched
`services, respectively. A new tunneling function is defined in
`the SGSN, and the GPRS LLC and BSSGP protocols have
`been modified in order to route 136MM signaling messages to
`the mobile station over the PDCH.
`Dual-mode mobile stations perform registration, authen-
`tication, authorization, and location update functions inde-
`pendently and in parallel with the ANSI-41 and GPRS
`networks while the mobile is camped on a PDCH. If an
`ANSI-41 SMS message needs to be delivered while a mobile
`
`is camped on a PDCH, it is delivered on the PDCH along
`with other packet data as tunneled LLC data. An incoming
`circuit call is indicated to a mobile with an ANSI-41 page
`message tunneled from the gateway MSC/VLR t o the
`mobile. The mobile moves to an IS-136 control channel to
`respond to the page and accept the incoming call. Similarly,
`to initiate a circuit call, the mobile stops any activity on the
`PDCH and moves to the IS-136 control channel. Circuit calls
`take precedence over packet data transactions. When a cir-
`cuit call is completed, the dual mode mobile station returns
`to camping on a PDCH.
`Half- Duplex Operation
`Half-duplex devices have complexity, size, and battery life
`benefits which make them attractive for applications that do
`not require full duplex (i.e., simultaneous transmission and
`reception) capability. Full-rate operation (i.e., similar to IS-
`136 voice) is half-duplex by definition because of fixed time
`offsets between the uplink and downlink [l]. The support of
`downlink double- and triple-rate (multislot) operation for
`packet data is quite straightforward, and is enabled by occa-
`sionally scheduling uplink time slots for obtaining automatic
`
`Figure 2. The GPRS-136packet data protocol stack.
`
`36
`
`IEEE Personal Communications June 1999
`
`IPR2018-1556
`HTC EX1016, Page 3
`
`

`

`Packet data channel (PDCH)
`
`Figure 3. Logical PDCHs.
`
`repeat request (ARQ) feedback. However, mobile
`stations cannot support more than full-rate operation
`on the uplink without simultaneous transmission and
`reception. GPRS-136 supports half-duplex operation
`on both the uplink and downlink for all mobile sta-
`tions regardless of their multislot (or bandwidth)
`capability. Double- or triple-rate half-duplex opera-
`tion on the uplink is made possible by carrying out a
`fixed allocation of a sequential number of time slots.
`No downlink transmissions are scheduled for the
`device during this period.
`Packet Data Channel Structure
`logical Channel Structure
`The PDCH consists of the following logical channels (Fig. 3):
`Packet broadcast control channel (PBCCH) for indicating
`generic system configuration related information
`Packet paging channel (PPCH) dedicated to delivering
`pages
`Downlink packet payload channel (PLCH) for delivering
`data generated by LLC, RRME, and GMM
`Packet channel feedback (PCF) for support of random
`access and reserved access on the uplink
`Uplink packet random access channel (PRACH) used by
`mobile stations to request packet data access to the system
`Uplink packet payload channel (PLCH) for delivering data
`generated by LLC and GMM
`Physical Channel Structure
`The PDCH uses 30 kHz R F channels and the time-slot struc-
`ture specified in IS-136 [1]. Each 40 ms frame on a 30 kHz
`R F channel consists of six time slots (three time slot pairs),
`numbered 1 to 6. One or more time slot pairs may be allocat-
`ed to a PDCH. The remaining time slot pairs may be allocat-
`ed to a digital control channel (DCCH) and/or digital traffic
`channel (DTC).
`
`Multirate Channels - A PDCH may be full-rate, double-rate,
`or triple-rate, depending on whether one, two, or three time
`slot pairs are allocated to the channel within each 40 ms
`frame. The channel bandwidth is indicated through system
`broadcast information.
`Superframe Structure - Sleep mode is defined for mobile sta-
`tions on the PDCH in order to improve mobile station standby
`time. A superframe structure similar to the IS-136 DCCH is
`defined on each full-rate portion of a PDCH in order to man-
`age sleep mode. The total number of time slots per superframe
`is 32. The superframe phase (SFP) is a modulo 32 up-counter
`which increments every 20 ms and helps mobile stations find
`the start of the superframe. If a double- or triple-rate PDCH is
`allocated on a particular frequency, superframe synchroniza-
`tion across the individual full-rate portions is required.
`
`Primary and Supplementary Phases - A multirate PDCH
`operates on a single channel frequency and consists of the fol-
`lowing:
`Primary phase
`Supplementary phase(s)
`In this context, a phase corresponds to a full-rate portion
`of a multirate PDCH. The primary phase always corresponds
`to a full-rate channel, and is the part of a multirate channel
`that contains logical broadcast and paging channels on the
`downlink. The supplementary phases on a multirate PDCH
`correspond to all time slots that are not part of the primary
`
`IEEE Personal Communications June 1999
`- . _ _
`
`~
`
`phase. The possible allocation of primary and supplementary
`phases is provided in Table 1.
`The primary phase superframe consists of time slots that
`are reserved for the PBCCH. Nominal paging time slots are
`determined from among the remaining time slots using a stan-
`dard hashing algorithm that relies on mobile station (sub-
`scriber) identity. The actual number of PPCH slots per
`superframe are configurable through a fast broadcast mes-
`sage. Except for the assignment of broadcast and paging chan-
`nels and SFP management, a mobile station views a double-
`or triple-rate PDCH as one common channel (i.e., a fat pipe)
`regarding data transmission or reception.
`
`Downlink and Uplink Burst Associations - Subchannels are
`defined on the uplink in order to allow sufficient processing
`time at both the mobile station and BS in conjunction with a
`random access event. The full-rate PDCH is defined to consist
`of three subchannels; there are six subchannels in each dou-
`ble-rate PDCH and nine subchannels in each triple-rate
`PDCH (Fig. 4).
`An enhanced packet channel feedback (PCF) field that relies
`on this association is defined on the downlink for uplink
`resource management. PCF flags are carried in downlink time
`slots to provide feedback for bursts sent previously on the uplink
`and to indicate subsequent assignments on the uplink. Mobile
`stations identify access opportunities on the uplink by reading
`PCF on the corresponding downlink time slots. The PCF field
`allows the support of contention access and reserved access on
`the same channel, and assumes different flag definitions depend-
`ing on the context. These flags are reliably encoded to ensure
`good performance even under poor channel conditions.
`For a full-rate PDCH, the uplink and downlink time slots
`are multiplexed to create three distinct access paths, as shown
`in Fig. 4. Assuming that path 1 (P1) in the downlink indicates
`that the next P1 time slot on the uplink is designated as a con-
`tention slot and is selected for a random access attempt, a
`mobile station sends the first burst of its access at that time
`
`Double-rate PDCH
`
`Double-rate PDCH
`
`1, 2, 4, 5
`2, 3. 5. 6
`
`1, II.
`
`2.5
`
`Double-rate PDCH
`
`184
`1, 3,4,6
`1, 2, 3, 4, 5, 6 1,4
`Triple-rate PDCH
`--
`W Table 1 . Multirate PDCHphase allocations.
`
`2,s
`3,6
`3,6
`2, 3. 5,6
`
`37
`
`IPR2018-1556
`HTC EX1016, Page 4
`
`

`

`Slot number
`
`subchannels
`
`~
`
`PI
`
`P2 P3
`
`W Figure 4. Downlink and uplink burst associations.
`
`(24.8 ms after receiving the full P1 slot on the downlink). The
`mobile station then starts reading the PCF flags in the next
`downlink P1 slot (21.8 ms after transmitting its burst) to
`determine the reception status of its initial access.
`Similarly, there are six access paths on each double-rate
`channel and nine access paths on each triple-rate channel. In
`general, a mobile station which has data to send transmits
`24.8 ms after obtaining an assignment via PCF on the down-
`link. The mobile station reads PCF on the downlink 21.8 ms
`after transmission in order to obtain the reception status of its
`transmitted burst.
`The Medium Access Control Function
`PDCH Selection, Reassignment, and Reselection
`Mobile stations are directed to a PDCH through IS-136 DCCH
`broadcast information. In cases where there are multiple
`PDCHs per sector, mobile stations are directed to a beacon
`PDCH. The beacon PDCH broadcast information indicates the
`number of PDCHs, as well as the bandwidth (full-, double-, or
`triple-rate) of each PDCH supported. Mobile stations then
`hash onto a particular PDCH depending on their identity and
`the number of PDCHs. Load balancing may be carried out by
`reassigning mobile stations across radio resources, and the per-
`formance may be further improved by maintaining MAC/RLP
`state across reassignments. Cell reselection procedures ensure
`continuity of service across cell boundaries since mobile sta-
`tions autonomously perform PDCH reselection when they
`detect a stronger signal from a neighbor cell.
`Active Mobile Identity Management
`Each mobile station is assigned a 7 bit temporary local identi-
`fier called an active mobile identity (AMI) which remains
`valid for one or several closely spaced transactions. The AMI
`
`is used to identify uplink time slot assignments and identify
`the recipient of data on the downlink. Of the 128 possible
`AMI values, only 89 are allowed for mobile stations engaged
`in point-to-point transactions. The all-zero AMI is excluded
`for PCF-related functions and also for identifying point-to-
`multipoint information on the downlink. The remaining 38
`AMIs closest to the all-zero codeword are excluded in order
`to increase PCF reliability.
`AMI assignment procedures are executed for both uplink
`and downlink transactions spanning more than one time slot.
`If a valid AMI has not already been assigned, the AMI assign-
`ment is carried out as a part of the transaction initiation pro-
`cedure. Once an AMI has been assigned to a mobile station,
`it is used for transactions in both directions (i.e., the AMI is
`assigned on the initiation of an uplink or downlink transac-
`tion, whichever begins first, and remains assigned until
`released). AMI release is based on the expiration of timers at
`both the mobile station and BS. The AMI release procedures
`are designed to ensure the following:
`The availability of the same AMI to the mobile station for
`
`CQF
`
`5
`5
`2
`Figure 5. Logical format of the PCF. The SLF is context-
`dependent and contains either an AMI for providing feedback
`to contention slots, or three fields for providing feedback to
`reserved slots.
`
`38
`
`JEEE Personal Communications June 1999
`
`IPR2018-1556
`HTC EX1016, Page 5
`
`

`

`another downlink or uplink transaction which is initiated
`before the expiration of the timer
`The release of the AMI first by the mobile station and sub-
`sequently by the BS in order to avoid any potential hijack-
`ing scenarios.
`
`assignment (SLA). The use of the SLF mechanism is context-
`dependent (i.e., the SLF field takes different values depend-
`ing on whether feedback is being provided for a contention
`slot or a reserved slot). This provides greater reliability for
`acknowledgments to reservation-based transmissions. In addi-
`tion, the same field may also be used for reliably reassigning
`the subchannel to the same user and providing continuous
`uplink channel quality feedback to the mobile station.
`Assignments to successful contending users or users sniff-
`ing for reservation-based access may be carried out by explic-
`itly identifying the user being assigned the next slot (i.e., by
`setting the SLA field to a valid coded active mobile identity).
`This approach is useful when there are several users on the
`channel, and subchannels (time slots) are constantly being
`assigned to new users. However, with fewer active users on
`the channel, subchannels may continuously be reserved for
`(reassigned to) the same users. Figure 6 shows that if there
`are three or fewer users on a full-rate channel, a round-robin
`assignment scheme is equivalent to reassigning subchannels to
`the same users. Similarly, Fig. 7 shows that if there are nine or
`fewer users on a triple-rate channel, a round-robin assignment
`scheme is equivalent to reserving one or more subchannel(s)
`for each user. In such cases, it is not necessary to use a coded
`AMI to reassign a subchannel to the same user. If there are
`fewer users on the channel, it is possible to provide a series of
`reservation-based transmission opportunities on each sub-
`channel through the use of a 1-bit CONTINUE/STOP (C/S)
`indicator. The CONTINUE indication is similar to BRI =
`BUSY [l], but is encoded more robustly. Since the CONTIN-
`UE (or BUSY) indication is associated with feedback for a
`reserved slot and is relevant only to the user who transmitted
`on the previous time slot on a particular subchannel, the C/S
`flag is included in the SLF field.
`PCF Encoding
`Subchannel Feedback - The 12-bit SLF field is used to
`acknowledge bursts transmitted in the previous slot. For feed-
`back corresponding to contention slots SLF is defined as fol-
`lows:
`SLF = valid coded AMI (provides an implicit ACK to the
`mobile with that coded AMI suggested or assigned)
`SLF = E-NAK (explicit NAK to all mobiles which attempt-
`ed access)
`The (12,7) code described below is used to encode this
`field. The all-zero codeword is reserved for indicating an
`explicit negative acknowledgment (E-NAK) to all mobiles
`which attempted contention access.
`For feedback corresponding to reserved slots, SLF is fur-
`ther divided into the following:
`R/N ( 5 bits): A (5,l) repetition code is used. A 1 indicates
`
`Transaction hitiation
`A new transaction is initiated by the transmit controller when
`a transmission opportunity is identified and if the transmit
`buffer contains new data. Downlink transactions may be
`acknowledged or unacknowledged, but uplink transactions are
`always acknowledged. The initiation of a new transaction is
`carried out through the transmission of a BEGIN PDU. The
`BEGIN PDU handshake (i.e., acknowledged transfer of the
`BEGIN PDU) is used to initialize the AMI and an RLP in
`either incremental redundancy (IR) or fixed coding (FC)
`mode for the transaction.
`For downlink transactions, the AMI and mode are assigned
`using the BEGIN PDU. For uplink transactions, the mobile
`station suggests an AMI and mode using the BEGIN PDU. If
`the suggested AMI and mode are acceptable to the base sta-
`tion, it provides an acknowledgment using the packet channel
`feedback (PCF) field. The PCF acknowledgment is treated as
`an implicit AMI and mode assignment by the mobile station.
`If the suggested AMI and/or mode are unacceptable to the
`base station, it provides a negative acknowledgment using
`PCF and subsequently assigns an AMI and/or mode for the
`transaction using a supervisory ARQ Status PDU.
`Packet Channel Feedback
`Functions - The IS-136 digital control channel [ l ] uses a
`shared channel feedback (SCF) field to provide acknowledg-
`ment and assignment functions. The SCF uses three flags -
`Receivedmot Received (R/N), Busy/Reserved/Idle (BRI), and
`a coded partial echo identifier (CPE) - in order to manage
`feedback and assignment functions. The DCCH procedures
`and fields as specified in IS-136 are not well suited for long
`packet data transactions.
`As described above, the uplink and downlink burst associa-
`tions on the PDCH result in three subchannels per full-rate
`channel, six subchannels per double-rate channel, and nine
`subchannels per triple-rate channel. A PCF field is associated
`with each subchannel, and 24 bits are used in each downlink
`time slot for the PCF. The PCF is designed to be more reli-
`able than the SCF and allows efficient management of con-
`tention access and reserved access on the same channel. For a
`particular subchannel, the PCF provides acknowledgment for
`a transmission on the previous time slot and also indicates
`assignment of the next time slot. Consistent canonical feed-
`back and assignment fields are defined as follows: acknowl-
`edgmenthegative acknowledgment (ACK/NAK) feedback
`identifying the transmitter when acknowledging a
`contention slot, and assignment of the next slot as
`idle (for contention-based access) or reserved (for
`access by a specific user). The approach present-
`ed simplifies the state machine since the acknowl-
`edgment for the previous slot and the assignment
`for the next slot are unambiguously identified
`with specific mobiles.
`Figure 5 shows the logical format of the PCF
`field associated with each subchannel. The main
`purpose of the PCF field is to acknowledge an
`access in the previous slot and to assign the next
`slot to a particular mobile. The acknowledgment
`and assignment functions are handled indepen-
`dently for each subchannel through two logical
`fields, subchannel feedback (SLF) and subchannel
`
`I
`
`I
`Figure 6. Round-robin assignments for users a, b, and c on a full-rate chan-
`nel. This is equivalent to reservation of subchannels 1, 2, and 3 for users a, b,
`and c, respectively.
`
`1.a 2,b 3,c 4.d 5.e 6.f 7.9 8.h 9.i
`
`l , a 2.b
`
`3.c 4,d 5,e 6,f 7,g
`
`W Figure 7. Round-robin assignments for users a, 6, . . ., i on a triple-rate chan-
`nel. This is equivalent to reservation of subchannels 1, 2, . . ., 9 for users a, b,
`. . ., i, respectively.
`
`IEEE Personal Communications June 1999
`
`39
`
`IPR2018-1556
`HTC EX1016, Page 6
`
`

`

`false error correction (i.e., interpre-
`tation of one codeword as another)
`may occur if two o r more errors
`occur in the channel. Two crror
`detecting capability is provided by
`increasing the minimum Hamming
`distance of the code to 4 through
`the addition of a single parity bit.
`The resulting (12,7) code may be
`viewed as an extended (1 1,7) code.
`
`, I
`
`Mobile Station Procedures -
`Mobile stations assume an appro-
`priate SLF structure depending on
`whether thcy are examining feed-
`back corrcsponding to a contention
`slot or a reserved slot. Thc dccod-
`ing rules at the mobile station pro-
`vide unequal error protection to
`the different flag values.
`A mobile which transmits in a particular con-
`tention slot reads the corresponding SLF field to
`determine if its transmission was successful:
`*On receiving an ACK (i.e., SLF/coded AMI
`match). it reads the SLA field on all subchan-
`nels it is capable of operating on to determine
`if it is granted a reservation.
`*On declaring a NAK (i.e., SLF/coded AMI mis-
`match), it follows random access procedures
`for attempting another contention access.
`A mobile which transmits in a particular reserved slot reads
`the corresponding SLF field to determine if its transmission
`was successful:
`On receiving an ACK (i.e., R/N = R), the mobile reads the
`C/S indicator to determine if it can transmit in the subse-
`quent slot.
`- On decoding C/S as CONTINUE, the mobile ignores the
`SLA field and assumes that it has permission to transmit in
`the subsequent time slot associated with the same subchan-
`nel.1
`- C/S = STOP implies that the mobile being acknowledged
`must continue sniffing (reading SLA) on that subchannel
`for reservation-based transmission opportunities.
`On receiving a NAK (i.e., R/N = N), the mobile reads the
`SLA field to determine if it has been assigned the subse-
`quent slot (Le., it tries to match SLA with its coded AMI)?
`Other active mobiles which are sniffing for reservations
`ignore the SLF field and examine the SLA-field for time slot
`assignments. With this scheme, it is possible to reserve sub-
`channels for some users and carry out round-robin assign-
`ments for other subchannels.
`Con tention Access
`Contention slots are provided on the uplink PDCH in order
`to allow mobile stations to initiate packet data transactions.
`The mobile station identifies a contention opportunity by
`reading the PCF. If the mobile station attempts transmission
`in a contention slot and is negatively acknowledged by the BS,
`it starts a timer and waits for an AMI and/or mode assign-
`ment from the BS. If no AMI/mode assignment is received
`and the timer expires, the mobile station must wait for T-Retiy
`idle (contention) slots before making another access attempt.
`The parameters used by the random access procedure are list-
`ed in Table 2.
`T-Retry is uniformly distributed in the closed interval [0,
`T-Retry-Max] where T-Retry-Max
`is a function of
`Access-Count, T-Retry-Init,
`and a. The parameter
`
`Table 2. Parameters for the random access procedure
`
`Table 3. Expressions for computing T-Retry-Max
`
`that the transmission was received (R), while a 0 indicates
`that the transmission was not received (N). The mobile
`declares R/N = R if the Hamming weight of the received 5-
`bit word is strictly greater than 3; otherwise, it declares R/N
`= N.
`C/S (5 bits): A (5,l) repetition code is used to encode this
`flag. A 1 indicates CONTINUE or C; that is, the mobile is
`assigned the subsequent time slot on the same sub-channel.
`A 0 indicates STOP or S; that is, the mobile must read the
`SLA field to determine subsequent assignments on that
`sub-channel. The mobile station declares C/S = C if the
`Hamming weight of the received 5-bit word is strictly
`greater than 2; otherwise, it declares C/S = S.
`Channel quality feedback (2 bits): This field provides feed-
`back on uplink channel quality to the mobile. A mobile
`capable of operation on multiple modulations can use this
`feedback to propose a different modulation €or subsequent
`reservation based transmission opportunities.
`
`Subchannel Assignment - The 12-bit SLA field can take the
`following values:
`SLA = valid coded AMI (assigns the subchannel to a
`mobile with that coded AMI)
`SLA = IDLE (identifies a contention opportunity)
`The (12,7) code described below is used to encode this
`field. The all-zero codeword is used as an IDLE indicator
`(i.e., to indicate a contention slot).
`
`AMI Encoding - Each mobile