`Exhibit 1015
`
`ZTE Corporation and ZTE (USA) Inc.
`ZTE Corporation and ZTE (USA) Inc.
`
`
`
`Inv. No. 337-TA-868
`RX-3029
`
`Evolution of UMTS Toward High-Speed
`Downlink Packet Access
`Arnab Das, Nandu Gopalakrishnan, Teck Hu, Farooq Khan,
`Ashok Rudrapatna, Ashwin 5ampath, Hsuandung 5u,
`5aid Tatesh, and Wenfeng Zhang
`
`An expanded effort is under way to support the evolution of the Universal
`Mobile Telecommunications System (UMTS) standard to meet the rapidly
`developin9 needs associated with wireless data applications. A new, shared
`channel--the high-speed downlink shared channel (HS-DSCH)--provides
`support to packet-switched high-speed data users. A number of
`performance-enhancing technologies are included in the high-speed
`downlink packet access (HSDPA) system to ensure high peak and average
`packet data rates while supporting circuit-switched voice and packet data
`on the same carrier, Lucent Technologies took a pivotal role in specifying
`many of these techniques, including adaptive modulation and coding (AMC),
`hybrid automatic repeat request (HARQ), and fat-pipe scheduling. In this
`paper, we provide system-level simulations results to indicate the achievable
`performance and capacity with these advanced technologies. We also discuss
`HSDPA protocol architecture along with the upfink and downfink control
`channel design and performance. We conclude with a discussion of potential
`enhancements for the future. © 2003 Lucent Technologies Inc.
`
`The deployment of third-generation (3G) mobile
`communication systems is under way with support
`for data rates up to 2 Mb!s (although data rates up to
`1.92 Mb/s are possible in Re199, realistic peak data
`rate for outdoor environments is about 384 kb/s).
`These data rates coupled with system latencies will
`not be sufficient to meet the increasing demands of
`data serxqces that ~e anticipated soon after 3G deploy-
`meat [7]. Therefore, extensive evolution programs in
`the slandards bodies of 3rd Generation Partnership
`Prqject (3GPP) and 3rd Generation Partnership Pm~ect
`2 {3 GPP2 ) were initiated to evolve 3 G systems beyond
`their basic capability in the first release.
`
`Traditionally, vmce communication has been
`
`the dominant application in wireless networks. As a
`
`result, cellular standards, such as Global System for
`
`Mobile Communication (GSM) and IS-95, were opti-
`
`mized for voice traffic only. With the recent explosive
`
`growth of the Internet, however, a need has arisen
`
`to offer both voice and reliable high-speed data access
`
`over wireless networks. Until recently, standardized
`
`3G systems such as CDMA2000* and Enhanced
`
`General Packet Radio Service (EGPRS) attempted to
`
`provide such capability by evo]ving the air interface of
`
`existing voice-centric second-generation (2G) systems.
`
`The service needs of voice and packet data, however,
`
`Bell Labs Technical Journal 7(3), 47-58 (2003) © 2003 Lucent Technologies Inc. Published by Wiley Periodicals, Inc.
`Published online in Wiley InterScience (wvvw.interscience.wiley.com). ¯ DOI: 10.1002ibltj.10018
`
`N K8681TC011070102
`ZTE Corporation and ZTE (USA) Inc.
`Exhibit 1015-00001
`
`
`
`Panel 1. Abbreviations, Acronyms, and Terms
`
`lxEV-DO--CDMA2000* evolution-data only
`lxEV-DV--CDMA2000 evolution-data/voice
`2G--second generation
`3G--third generation
`A21R--asynchronous adaptive incremental
`redundancy
`AAIR--asynchronous and adaptive incremental
`redundancy
`ACK--acknowledgment
`AMC--adaptive modulation and coding
`APP--a posteriori probability
`ARQ--automatic repeat request
`CQI--channel quality indication
`CRC--cyclic redundancy check
`CRNC--controlling RNC
`DL--downlink
`DPCCH--dedicated physical control
`channel
`DPDCH--dedicated physical data channel
`DSCH--downlink shared channel
`DTX--discontinuous transmission
`FCS--fast cell selection
`FER~frame error rate
`HARQ--hybrid ARQ
`HSDPA--high-speed downlink packet access
`HS-DPCCH--high-speed dedicated physical
`control channel
`
`HS-DSCH--high-speed downlink shared channel
`HS-SCCH--high-speed shared control channel
`IP--Internet protocol
`MAC-c--MAC-control
`MAC-hs--MAC-high speed
`MAC--medium access control
`MCS--modulation and coding scheme
`MIMO--multiple input, multiple output
`MMSE--minimum mean squared error
`NACK--negative acknowledgment
`NAIR--non-adaptive incremental redundancy
`NCCS--network controlled cell selection
`RAN--radio access network
`RNC--radio network controller
`SFN--system frame number
`SIN R--signal-to-interference-plus-noise ratio
`SPARCS--Synergistic Power And Rate Control
`System
`TBS--transport block size
`TFRC--transport format and resource control
`TF--transport format
`TTI--transmission time interval
`UL--uplink
`UMTS--Universal Mobile Telecommunications
`System
`UTRAN--UMTS terrestrial and radio access
`network
`
`are different (e.g., low latency and no jitter for
`
`carrier. Hence, an expanded effort is under way in
`
`isochronous bidirectional streams such as voice
`
`3GPP and 3GPP2 for the evolution of UMTS and
`
`contrasted with modest latencies and jitter for packet
`
`CDMA2000 1 X, respectively. These 3G evolutions--
`
`data, resilience of voice for low frame errors con-
`
`high-speed downlink packet access (HSDPA) and
`
`trasted with extremely low error rates for data
`
`lxEV-DV--address the challenge of supporting the
`
`applications). Not surprisingly, the support of delay-
`
`separate and often conflicting needs of voice and
`
`tolerant data services in these standards proved to be
`
`high-speed data simultaneously and efficiently on the
`
`inadequate because voice-centric techniques were
`
`same carrier in a manner that is fully backward com-
`
`applied to do resource allocation for packet data.
`
`patible. To address these needs and issues, Lucent
`
`The recently standardized CDMA2000 lxEV-I)O sup-
`
`Technologies developed its Synergistic Power and Rate
`
`ports efficient packet data service over a dedicated
`
`Control System (SPARCS). Several key technologies
`
`CD~’VLA2000 lX carrier by using a design philosophy
`
`were developed and adapted to address the differing
`
`that is markedly different from that of CDMA2000
`
`system requirements of HSDPA and lxEV-DV. This
`
`and EGPRS, thereby resulting in a far superior
`
`adaptation has required significant system-specific in-
`
`performance. However, lxEV-DO is not backward
`
`novation that Lucent spearheaded and successfully
`
`compatible with existing lX systems and, more im-
`
`championed in the standards bodies. These technolo-
`
`portantly, does not support voice-service on the same
`
`gies are described briefly in the section below.
`
`48
`
`Bell LabsTechnical Journal
`
`N K8681TCO 11070103
`
`ZTE Corporation and ZTE (USA) Inc.
`Exhibit 1015-00002
`
`
`
`Ips
`
`I~adio interface! ’ ~Access network (UTRAN)
`
`Core network
`
`’, External networks
`
`3G -Third generation
`GPRS- General Packet Radio Service
`HLR - Home location register
`ISDN - Integrated services digital network
`MSC - Mobile switching center
`PSTN - Public switched telephone network
`
`- Radio network controller
`RNC
`-Serving GPRS support node
`SGSN
`-User equipment
`UE
`- Universal Mobile Telecommunications System
`UMTS
`UTRAN - U MTS terrestrial rad io access network
`
`Figure 1.
`Simplified UMTS network architecture.
`
`In 3GPP, different areas of evolution of the UMTS
`
`system are currently under way. HSDPA is one step in
`
`the evolution of UMTS aimed at optimizing the air
`
`interface to support higher data rates up to a data rate
`
`of 10 Mb/s. Besides the substantial increase in peak
`
`data rates, the objectives of HSDPA are to achieve a
`
`reduction in system delays and thereby increase
`
`system capacity and throughput on the downlink.
`
`HSDPA study has shown that, using a single UMTS
`
`carrier, a 10.8-Mb/s peak rate is achievable in
`
`the downlink, which would significantly increase the
`
`downlink packet access speed over the current air in-
`
`terface. This is achieved by implementing a number of
`
`new physical layer attributes such as adaptive modu-
`
`lation schemes, fast channel state feedback from user
`
`equipment (UE; i.e., mobiles), flexible and dyna~nic
`
`scheduling, and hybrid automatic repeat request
`
`(HARQ) channel coding--all within a new suitable
`
`architecture. These schemes allow fast link adapta-
`
`tion by selecting appropriate modulation size, number
`
`of codes, and the rate of the channel encoder to track
`
`variations of the radio channel. YVhile HSDPA is opti-
`
`mized mainly for low mobility urban environments, it
`
`will operate well also in other environments with
`
`higher mobile speeds. An important consideration in
`
`the evolution path of UN{TS toward HSDPA was to
`
`provide graceful migration for the operators from
`
`Re199/4 to HSDPA capable networks with minimal
`impacts and costs.
`A high-level view of the UMTS architecture is
`shown in Figure 1. It comprises a radio access net-
`work part, UMTS terrestrial radio access network
`(UTRAN), that can interface to a variety of core net-
`works. The core networks contain mobile switching
`centers and gateways to various circuit and packet net-
`works. UTRAN is linked to the core network via back-
`haul facilities, for example. TIlE 1, STM-x. UTRAN
`itself comprises cell sites called Node Bs that contain
`the radio transceivers and radio network controllers
`
`(RNCs). Several Node Bs interface with an RNC where,
`in addition to call setup and control activit}; tasks such
`as radio resource management and frame selection in
`soft handoff are carried out. Node Bs and RNCs are
`connected via links that use ATM-based packet
`transport. UMTS Re199/4 defines a downfink shared
`channel that can time multiplex packet data users.
`However, the scheduling and resource management is
`performed at the RNC, thereby incurring large delays
`on the Iub interface. The downlink shared channel
`(DSCH), therefore, is not agile enough to provide very
`high data rates through dynamic scheduling and rate
`selection as desired and as outlined here.
`
`For HSDPA, a new channel--the high-speed down-
`link shared channel (HS-DSCH)--is defined and is
`
`Bell Labs Technical Journal 49
`
`N K8681TCO 11070104
`
`ZTE Corporation and ZTE (USA) Inc.
`Exhibit 1015-00003
`
`
`
`terminated in the Node B in the UTRAN. This is
`unlike in Re19914 in which the corresponding chan-
`
`nel, the DSCH, terminates at the RNC. In addition to
`enhancements to the air interface, other system
`capacity enhancements are being developed within
`both the standards arena and Lucent. Improvements
`will cover UTRAN and the core network and areas
`such as distributed scheduler and Internet protocol
`
`radio access network (IP-RAN).
`
`The key technologies nsed in HSDPA consist of
`fast scheduling, adaptive modulation and coding
`(AMC), and HARQ. These technologies along with
`the advanced channel structure and radio resource
`control in the Node B (as opposed to the RNC)
`ilnprove system capacity by a factor of greater than 2
`over Re199 UMTS.
`
`Fast Scheduling
`In recent years, considerable work has been done
`in the area of multi-user transmission on fading chan-
`nels. Information theoretic results demonstrate that
`the total downlink information capacity (in the in-
`formation theoretic sense) is achieved through chan-
`nel quality sensitive scheduling. In other words, sector
`throughput is maximized by the scheme in which the
`base station assigns resources to one user or a subset
`of users at a given time based on their channel qual-
`ities. ~Zhile traditional forms of diversity in wireless
`systems include time, frequency, and antennas, such
`scheduling provides diversity that arises from inde-
`pendent fading channels across different users.
`Scheduling in conjunction with AMC then allows the
`base station to select the user with the best channel
`quality and the best-suited modulation and coding
`
`scheme (MCS) for that user at the time. Thus, multi-
`user diversity takes advantage of rather than com-
`pensates for channel fading. The sector throughput
`increases monotonically with the number of users.
`
`Adaptive Modulation and Coding
`The benefits of adapting the transmission para-
`meters in a wireless system, especially a CDMA
`system, to the changing channel conditions are well
`known. Fast power control is one such example that
`is critical for CDMA systems with voice users. In the
`
`50
`
`Bell LabsTechnical Journal
`
`wireless data context, higher data rates can be
`achieved by varying the inodulation level and/or the
`channel coding rate appropriately based on estimating
`the channel quality. In a system with AMC, users in
`favorable positions or users experiencing an "up fade"
`typically will be assigned higher order modulation and
`higher code rates. This represents a paradigm shift to
`rate control rather than power control for wireless
`data. An added benefit to keeping the power constant
`
`is that the inter-cell interference variations on the
`downlink are reduced.
`HARQ
`Link adaptation via AMC suffers degradation from
`a few sources. First, AMC provides limited granularity
`in data rate selection, and often the channel quality
`
`estimates dictate a rate that is in between two allowed
`MCSs. Second, estimates of link quality are prone to
`error due to delay between the time of measurement
`
`and the time of rate selection and also dne to
`measurement error. HARQ provides some level of
`robustness through fast retransmissions at the physi-
`cal layer. Retransmitted copies are combined at the
`receiver and then decoding is attempted again. The
`HARQ scheme in HSDPA is based on Lucent’s proposal
`of asynchronous adaptive incremental redundancy
`
`(A2IR) [3]. In A2IR, the retransmissions can be sched-
`uled exactly like original transmissions. Moreover,
`retransmissions can use a different nmnber of chan-
`nelization codes and modulation and coding rates
`than the original transmission.
`
`In order to support fast scheduling, adaptive mod-
`ulation and coding, and HARQ, a new medium access
`layer called MAC-hs is introduced in the Node B.
`
`Moreover, some ne~v control channels both on the
`downlink and uplink are introduced.
`
`MAC-hs Architecture
`UMTS supports a wide range of data rates for
`services and variable bit-rate operations for a given
`service. It also allows efficient multiplexing of mnlti-
`ple logical data streams to a user. Many of these ac-
`tivities are accomplished at the medium access control
`(MAC) layer that resides in the RNC. For example,
`the MAC layer performs dynamic data rate selection
`
`N K8681TCO 11070105
`
`ZTE Corporation and ZTE (USA) Inc.
`Exhibit 1015-00004
`
`
`
`for a service from within a set. This is accomplished by
`defining a basic unit of time--the transmission time
`interval (TrI)--and selecting an appropriate transport
`format (TF) for each TTI. Among other things, the TF
`determines the number of information bits that will be
`transferred during the TTI and, thereby, the current
`data rate for the service. While Re199/4 supports a
`shortest TTI of 10 ms, in HSDPA, due to the agility
`required to exploit fast changing channel conditions,
`
`the TTI is shortened to a value of 2 ms.
`For HSDPA, the following functionalities have to
`be added to the existing MAC layer architecture:
`¯ The inclusion of new functionality for HARQ and
`HSDPA scheduling.
`
`° The provisioning of fast UE channel feedback and
`fast scheduling or resource allocation thai allows
`the base station (Node B) to take advantage of the
`good channel conditions of the mobiles CUE).
`° A new" MAC layer, MAC-hs, located at the Node B
`in addition to a new channel, the HS-DSCH, that
`is defined for HSDPA. Relocating the MAC-hs
`to the Node B facilitates fast scheduling by avoid-
`ing the latency involved when the MAC-hs is
`placed at the RNC.
`The entities within the MAC-hs for a UE, illus-
`trated in Figure 2, consist of the flow control, the
`scheduling and priority handling, HARQ, and the
`transport-format and resource-related information
`
`to MAC-c/sh or MAC-d
`
`Scheduling/priority handlingI
`
`trol
`
`-
`
`, [
`
`Associated uplink
`signalling
`
`HS-DSCH
`
`Associated downlink
`signalling
`
`-Hybrid automatic repeat request
`
`HARQ
`- Medium access control
`MAC
`- MAC-control
`MAC-c
`MAC-dsh- MAC-control/shared
`MAC-d - MAC-dedicated
`MAC-hs -MAC-high speed
`TFRC -Transport format and resource control
`
`Figure 2.
`MA C-hs architecture.
`
`Bell Labs Technical Journal 51
`
`N K8681TCO 11070106
`
`ZTE Corporation and ZTE (USA) Inc.
`Exhibit 1015-00005
`
`
`
`(TFRI) entities. Motivated by the need to perform
`
`HARQ operation at the Node B level, the MAC-hs con-
`
`tains the HARQ engine or entity that supports multi-
`
`ple instances of stop-and-wait HARQ processes. At any
`
`TTI, the scheduler in the ~MAC-hs can transmit a max-
`
`imum of one code block from a single HARQ process
`
`from a single UE--i.e., there is only one HARQ process
`
`per TTI per UE. With code division multiplexing
`
`(CDM), more than one UE can simultaneously receive
`
`transmissions in a single TTI.
`
`The data for multiple priority classes is received
`
`from the Iub frame protocol at the MAC-hs. The
`
`scheduler receives this information from the lub
`
`frame protocol and uses it in making scheduling de-
`
`cisions at the scheduling and priority handling entity.
`
`This function manages HS-DSCH resources between
`
`HARQ entities and data flows according to their pri-
`
`ority class. Based on status reports from associated
`
`uplink signaling, it determines whether to send a new
`
`transmission or a retransmission. It also sets the pri-
`
`ority class identifier and the sequence number for
`
`each new data block being serviced.
`
`A flow control mechanism is defined in order to
`
`control the amount of data that the RNC can forward to
`
`the Node B. This fm~ctionis intended to limit layer 2 sig-
`
`naling latency and reduce discarded and retransmitted
`
`data as a result of HS-DSCII congestion. The flow con-
`
`trol is provided independently per priority class.
`
`SF
`SF
`SF
`SF
`SF
`
`1
`2
`4
`8
`16
`
`Physical channels (codes) to
`which HS-DSCH is mapped
`
`SFHsDPA 16 (example)
`Number of codes to which HSDPA transmission
`is mapped: 12 (example)
`
`HS-DSCH - High-speed downlink shared channel
`HSDPA -High-speed downlink packet access
`
`Figure 3.
`HSDPA mapping to physical channels with fixed
`spreading factor.
`
`2.0 ms
`
`All codes to
`which HSDPA
`transmission is
`mapped (5 in
`this example)
`
`~ Data to UE #1 ~ Data to UE #2
`
`Data toUE#3
`
`- User equipment
`UE
`HSDPA - High-speed downlink packet access
`
`Figure 4.
`Sharing by means of time multiplexing as weft as code
`multiplexing.
`
`The last remaining functional entity in the MAC-hs
`
`feedback information (i.e., acknowledgment/negative
`
`is the transport format and resource control (TFRC)
`
`acknowledgment [ACKiNACK] ) to the Node B, a new
`
`selection emity. Its function is selection of an appro-
`
`channel, the high-speed dedicated physical control channel
`
`priate transport format and resource combination (i.e.,
`
`(HS-DPCCH), is defined in the uplink.
`
`data rate and the physical layer attributes required to
`
`tlS-DSCH. The HS-DSCH preferably supports
`
`achieve it) for the data to be transmitted on HS-DSCH
`
`one UE at a time. However, more than one UE can be
`
`transport channel. The TFRC information is sent in the
`
`code-multiplexed within a TTI if a backlog from the
`
`downlink high-speed shared control channels (HS-SCCHs).
`
`single user cannot fill all the available power and codes
`
`More than one UE, each with its own TFRCs, can be
`
`within a TTI and/or the UE capability do not allow use
`
`supported in an HSDPA TTI via CDM.
`
`of all the available channelization codes. The physical
`
`HSDPA Channel Structure
`
`channels to which HS-DSCH is mapped has a fixed
`
`spreading factor of 16, as shown in Figure 3. The phys-
`
`The HSDPA service is carried over a new chan-
`
`ical channels to which HS-DSCH is mapped can still be
`
`nel, HS-DSCH, which is terminated in the Node B.
`
`shared between "users" in the time domain as well as in
`
`Moreover, a new downlink shared control channel is
`
`the code domain, as shown in Figure 4. A physical
`
`defined to carry the scheduling and HARQ informa-
`
`layer block diagram conceptually showing the trans-
`
`tion to the scheduled UE. In order to provide HARQ
`
`mit chain for this approach is depicted in Figure ~.
`
`52
`
`Bell iabsTechnical Journal
`
`N K8681TCO 11070107
`
`ZTE Corporation and ZTE (USA) Inc.
`Exhibit 1015-00006
`
`
`
`N Transport
`Blocks
`
`W~sF
`
`AMC- Adaptive medulation and coding
`PSK - Phase shift keying
`
`QAM - Quadrature amplitude modulation
`QPSK - Quadrature PSK
`
`Figure 5.
`HSDPA physical layer structure.
`
`HS-SCCH. The adaptive nature of the HS-DSCH
`provides ample flexibility in radio resource allocation.
`This flexibility is obtained at the price of fast control
`signaling. Precisely, for a UE to successfully receive a
`
`transmission on the HS-DSCH, the TFRI and the
`HARQ-related information must be delivered to the UE
`before the transmission takes place. The detailed
`control message itself can take up a large amount of
`downlink bandwidth. Fortunately, due to fat-pipe
`scheduling of the HS-DSCH, only the few UEs sched-
`uled at a given moment need the control information.
`Thus, similar to the downlink data packets, the control
`messages for different UEs are transmitted, when the
`UEs are scheduled, on a limited number of HS-SCCHs.
`For each HS-DSCH, the number of associaled
`HS-SCCHs can range from a minimum of one to a
`maximum of four. The UE has the capability to simul-
`taneously monitor four HS-SCCHs. For each HS-DSCH
`
`TTI, each HS-SCCH carries HS-DSCH-related down-
`link signaling for one UE. The downlink signaling mes-
`sage contains TERI (including channelization code set,
`modulation scheme, and transport-block size) and
`HARQ information (including HARQ process number,
`redundancy version, new data indicator, and UE ID).
`To have the best adaptability to the channel condition,
`the HS-DSCH transmission should be as early as
`possible after the channel quality feedback is received.
`On the other hand, sending the signaling message
`before the HS-DSCH TTI helps the UE tune to the
`correct channelization codes and modulation scheme
`
`371Tslot (2 ms)
`¯
`HS-SCCH L
`
`HS-DSCH
`
`2 ~<Tslot
`
`J
`HS-DSCH~FI (2 ms)
`L
`J
`
`HS-DSCH- High-speed downlink shared channel
`TTI
`-Transmission time interval
`
`Figure 6.
`Tirning structure for HS-DSCH control signaling.
`
`in advance and avoid a huge amount of buffering. As
`
`a result, a tradeoff was made to stagger the HS-SCCH
`
`and the HS-DSCH (Figure 6). Once a UE is scheduled,
`
`the signaling message is sent immediately on the
`
`HS-SCCH and power controlled toward that UE.
`
`The HS-DSCH TTI for the UE is sent two slots after the
`
`start of the corresponding HS-SCCH TTI, allowing the
`
`UE two slots of time to decode the time critical infor-
`
`mation, namely, the channelization code set and the
`
`modulation scheme used. This not only reduces
`
`the buffer requirement of the UE, but also lets the UE
`
`start decoding the HS-DSCH earlier.
`
`The HS-SCCH structure is shown in Figure ~.
`
`The signaling message is divided into two parts, with
`
`part 1 containing the time critical information on chan-
`
`nelization code set and modulation scheme, and part 2
`
`consisting of transport block size and HARQ-related
`
`Bell Labs Technical Journal 53
`
`N K8681TCO 11070108
`
`ZTE Corporation and ZTE (USA) Inc.
`Exhibit 1015-00007
`
`
`
`Part 1
`
`Part 2
`
`16
`CRC + UE ID
`
`1/3 convolutional coding
`
`ID -Identification
`
`UE- User equipment
`
`UE ID based
`scrambling sequence
`
`Rate matching
`g
`
`Figure &
`Scrambling of part I by UE-specific ID.
`
`Slot 1
`
`Slot 2 + Slot 3
`
`CCS -Channelization code set
`CRC - Cyclic redundancy check
`HARQ - Hybrid automatic repeat request
`ID - Identification
`MOD - Modulation
`TBS
`-Transport block size
`- User equipment
`UE
`
`Figure 7.
`HS-S CCH structure.
`
`information. A 16-bit UE-specific cyclic redundancy
`check (CRC) is computed over part 1 and part 2 and
`attached to part 2. The UE ID is implicitly included in
`the CRC byusing the UE ID as the starting state of the
`shift register, or using an all-zero starting state, and
`then masking the computed CRC with the UE ID.
`These two methods are equivalent. The two parts
`are further attached with tail bits, convolutionally
`
`encoded, and then transmitted in slot 1 and slots 2 and
`3, respectively.
`At every HS-SCCH TTI, the UE monitors all four
`HS-SCCHs and tries to extract part 1 information from
`them. Although there is aUl~-specific CRC at the end of
`part 2 to prevenl false alarm for the unintended UE,
`the UE processing resource is wasted due to buffering of
`all four HS-SCCHs. This waste in resource is prevented
`by scrambling the post convolutionally encoded part 1
`with UE-specific ID (Figure 8). At the receiver, after
`descrambling by the UE specific ID, a suitable decoder
`metric (e.g., Viterbi or Yamamoto-Itoh) may be used
`to determine if the part 1 information was intended
`for the UE or not. The descrambling, decoding, and
`
`54
`
`Bell [absTechnical Journal
`
`validation of part 1 information should be completed
`before the start of the corresponding HS-DSCH.
`As the HS-SCCH carries the signaling message
`for the HS-DSCH, clearly, successful operation of
`the HS-DSCH relies on a low HS-SCCH error rate.
`Unfortunately, unlike the HS-DSCH that can benefit
`from the AMC and HARQ, the HS-SCCH has a fixed
`transmission rate and does not allow retransmissions.
`Although the HS-SCCH is power controlled, its ex-
`tremely short frame size (2 ms) provides little time
`diversity. With the presence of fading and channel qual-
`ity feedback delay, it turns out that the power required
`to guarantee a certain HS-SCCH frame error rate, say
`1%, is quite large. Since the HS-SCCH is sharing power
`with the HS-DSCH and other dedicated channels, this
`incurs a significant loss in the system capacity. The
`power consumption m~d margin of the HS-SCCH can
`be reduced by considering transmit and!or receive
`diversity. Between the two categories of transmit
`diversity--open-loop and dosed-loop--open-loop
`provides performance gain over the single-antenna sys-
`tem at all mobile speeds. The closed-loop transmit di-
`versity schemes outperform the open-loop schemes at
`low mobile speed where the feedback rate can track
`channel variations. At high mobile speed where the
`feedback loop fails to track the channel variation,
`the closed-loop schemes provide little gain over the
`single- antenna system and are no longer better than
`the open-loop schemes. Considering the additional
`channel condition feedback required for the closed-
`loop schemes and the robustness under variable mobile
`speeds, open-loop transmit diversity schemes are pre-
`ferred over the closed-loop diversity schemes.
`
`N K8681TC011070109
`
`ZTE Corporation and ZTE (USA) Inc.
`Exhibit 1015-00008
`
`
`
`1 TTI (subframe)
`
`1 slot
`
`ACK - Acknowledgment
`CQI - Channel quality indication
`
`NACK - Negative acknowledgment
`TTI -Transmission time interval
`
`Figure 9.
`HS-DPCCH field structure.
`
`Table I. Basis sequences for (20, 5) code.
`
`I
`
`0
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`Mi,0
`
`Mi,1
`
`ii,2
`
`Mi,3
`
`Mi,4
`
`1
`
`0
`
`1
`
`0
`
`1
`
`0
`
`1
`
`0
`
`1
`
`0
`
`0
`
`1
`
`1
`
`0
`
`0
`
`1
`
`1
`
`0
`
`0
`
`1
`
`0
`
`0
`
`0
`
`1
`
`1
`
`1
`
`1
`
`0
`
`0
`
`0
`
`0
`
`0
`
`0
`
`0
`
`0
`
`0
`
`0
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`The 8 information bits of part 1 and 13 infor-
`mation bits of part 2 are appended with a 16-bit CRC.
`The generator polynomial for this CRC is chosen to
`be the same as the one defined in Re199 UMTS.
`Additionally, in order to make the information in the
`HS-SCCH specific to a given UE that is scheduled
`on the HS-DSCH, the CRC is made UE specific by
`adding the UE ID modulo two to the 16-bit CRC.
`HS-DPCCtt. The HS-DPCCH is a new uplink phys-
`ical channel used by every Release 5-capable UE in
`order to supporl HSDPA operation on lhe downlink.
`As defined currently, it is used to carry two pieces of
`feedback information--channel quality indication
`(CQI) and HARQ ACK/NACK. Physically, it comprises
`a 256-ary (spreading factor) channelization code
`
`framed over three slots (one HSDPA TTI three
`slots 2 ms) comprising two fields for the
`
`ACK/NACK and CQI This channel is code multi-
`plexed along with other uplink physical channels and
`
`is carried on either the I or the Q sub-carrier, de-
`pending on whether an uplink dedicated physica! data
`
`cham~el (UL DPDCH) also exists.
`Figure 9 schematically illustrates the current
`working assumption of the structure of the HS-
`
`DPCCH sub-frame (or TTI). The ACKiNACK field
`occupies the first slot worth 10 channel bits, and
`hence it repeats the one information bit of ACK or
`
`NACK ten times. The second and third slots, worth
`20 channel bits, are used to carry coded CQI infor-
`mation represented by 5 information bits. A (20, 5)
`block code whose code words are a linear combina-
`tion of the 5 basis sequences denoted Mira defined in
`
`Table I is used to code the CQI information.
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`1
`
`0
`
`1
`
`0
`
`1
`
`0
`
`0
`
`0
`
`0
`
`0
`
`1
`
`0
`
`0
`
`1
`
`1
`
`0
`
`0
`
`0
`
`0
`
`0
`
`0
`
`1
`
`1
`
`1
`
`1
`
`0
`
`0
`
`0
`
`0
`
`0
`
`1
`
`1
`
`1
`
`1
`
`1
`
`0
`
`0
`
`0
`
`0
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`1
`
`0
`
`The ACK/NACK bit is used to signal in a fast man-
`ner to the Node B by the UE whether the HS-DSCH
`transmission intended for it was successful or not.
`Latency is minimized by use of dedicated physical
`layer resources only for the ACK/NACK creation,
`
`Bell Labs Technical Journal 55
`
`N K8681TC011070110
`
`ZTE Corporation and ZTE (USA) Inc.
`Exhibit 1015-00009
`
`
`
`Tslot (.67 ms)
`
`Uplink DPCCH
`
`~
`
`3×Tslot (2 ms)
`
`HS-DSCH at UE
`
`Uplink HS-DPCCH QI A/NA[-~A/NA QI A/NA QI ~
`
`~’UEP (7.5 XTslot 5 ms) 0-255~ chips~
`
`m×256 chips
`
`A -Acknowledgment
`HS-DPCCH - High-speed dedicated physical control channel
`NA - Negative acknowledgment
`
`QI - Quality indication
`UE - User equipment
`
`Figure 10.
`Timing structure at UE for UL HS-DPCCH control signaling.
`
`turnaround, and processing operations. Latency min-
`
`processing functions with the ACK/NACK bit and (2.5
`
`imization is critical to minimize the number of HARQ
`
`slots propagation delay HS-DPCCH offset) for
`
`processes spawned for a given UE (that adds to
`
`Node B to perform scheduling and signal processing
`
`overhead) in order to fully utilize the channel and,
`
`functions with the C QI information that immediately
`
`more importantly, to minimize the packet delay. This
`
`follows the ACK/NACK bit. The assumption under-
`
`latency is essentially fixed across all UEs regardless of
`
`lying the last sentence is that six TTIs (18 slots) is the
`
`their capability. This leads to a synchronous HSDPA
`
`minimum separation between retransmissions of
`
`uplink even though downlink scheduling of the
`
`the same HARQ process on the HS-DSCH. Of course,
`
`HS-DSCH for this UE may be asynchronous for new
`
`the actual separation can be larger depending on the
`
`transmissions as well as retransmissions. Synchronous
`
`asynchronous scheduler operation for HARQ retrans-
`
`uplink has the advantage of lower uplink overhead as
`
`missions and/or the processing time of the actual Node
`
`well as more robustness against misinterpretation by
`
`B implementation. Additionally, the maximum num-
`
`the Node B of the ACK/NACK feedback.
`
`ber of HARQ processes possible per UI~ is eight, due to
`
`F~gure IO shows the timing offset between the
`
`the three bits overhead allocated on the HS-SCCH for
`
`uplink HS-DPCCH and the uplink DPCCH with respect
`
`indexing this parameter.
`
`to the downlink HS-DSCH. The code-multiplexed up-
`
`The logic used for ACK/NACK/DTX signaling by
`
`link HS-DPCCH start~ m X 256 chips after the start
`
`the UE is as follows:
`
`of the uplink DPCCH with m selected by the UE such
`
`An ACK (0 bit) is signaled if the UE-specific CRC
`
`that the ACKiNACK transmission (of duration 1
`
`in part 2 of one of the monitored HS-SCCHs passes
`
`timeslo~) commences within ~he first 0 255 chips after
`
`the check (implying that this UE considered itself
`
`7.5 slots following the end of the received HS-DSCH.
`
`scheduled) and the data-specific CRC check passes
`
`The UE processing time is therefore maintained at
`
`on the associated HS-PDSCH (implying that the
`
`7.5 slots (5.0 ms) as the offset between DPCCH and
`
`packet was successfully decoded requiring no
`
`HS-DPCCH varies. The ACK bit is sent on the first slot
`
`of the code-multiplexed uplink H$-DPCCH. This
`
`further retransmissions).
`¯ A NACK (1 bit) is signaled if the UE-specific CRC
`
`leaves approximately 4.5 slots-512 chips (propagation
`
`in part 2 of one of