wcdma and cdma2000 - the radio interfaces for future

WCDMA and cdma2000 - The Radio Interfaces for Future Mobile Multimedia Communications - Part IIDecember 15, 2001 1

WCDMA and cdma2000 - The Radio Interfaces for Future Mobile Multimedia

Communications - Part II

Emre A. Yavuz

6.0 WCDMA

The WCDMA scheme has been developed as a joint effort between ETSI and ARIB dur-ing the second half of 1997 [30]. The ETSI WCDMA scheme and the ARIB WCDMAscheme have been developed from the FMA2 scheme in Europe [3137] and the Core-A

scheme in Japan [3843], respectively. The uplink of the WCDMA scheme is based

mainly on the FMA2 scheme, while the downlink is based on the Core-A scheme. In this

section, main technical features of the ARIB/ETSI WCDMA scheme are presented. Table

2 lists the main parameters of WCDMA [2].

TABLE 2. Parameters of WCDMA

Channel bandwidth 1.25, 5, 10, 20 MHz

Downlink RF channel structure Direct spread

Chip rate (1.024)a/4.096/8.192/16.384 Mc/s

Roll-off factor for chip shaping 0.22

Frame length 10 ms/20 ms (optional)

Spreading modulation Balanced QPSK (downlink)

Dual channel QPSK (uplink)

Complex spreading circuit

Data modulation QPSK (downlink)

BPSK (uplink)

Coherent detection User dedicated time-multiplexed pilot (downlink and uplink); no common pilot in downlink

Channel multiplexing in uplink Control and pilot channel time multiplexed

I & Q multiplexing for data and control channel

Multirate Variable spreading and multicode

Spreading factors 4256

Power control Open and fast closed loop (1.6 kHz)

Spreading (downlink) Variable length orthogonal sequences for channel

seperation Gold sequences 218 for cell and user

seperation (truncated cycle 10 ms)


6.1 Carrier Spacing and Deployment Scenarios

The carrier spacing has a raster of 200 kHz and can vary from 4.2 to 5.4 MHz. The differ-ent carrier spacings can be used to obtain suitable adjacent channel protections dependingon the interference scenario. Figure 15 shows an example for the operator bandwidth of 15MHz with three cell layers. Larger carrier spacing can be applied between operators thanwithin one operator’s band in order to avoid inter-operator interference. Interfrequencymeasurements and handovers are supported by WCDMA to utilize several cell layers andcarriers.

FIGURE 15. Frequency utilization with WCDMA

6.2 Radio - Interface Protocol Architecture

The general system architecture of UMTS/IMT-2000 includes user equipment (MS),UMTS terreterial radio-access network (UTRAN) and a core network. The functional lay-ering of this system introduces the concepts of access stratum and nonaccess stratum [44].

Access Stratum, is a 3GPP term used to identify the protocol layers directly involved ininteractions between the infrastructure and the subscriber equipment. Typically, itincludes the Radio Resource Control sublayer, Layer 2 and Layer 1. The corresponding3GPP2 term is Lower Layers.

Spreading (uplink) Variable length orthogonal sequences for channel

seperation Gold sequences 241 for user seperation (different time shifts in I and Q channel, truncated

cycle 10 ms)

Handover Soft handover, Interfrequency handover

aIn the ARIB WCDMA proposal a chip rate of 1.024 Mc/s has been specified,

whereas in the ETSI WCDMA has not.

TABLE 2. Parameters of WCDMA


Nonaccess Stratum, is a 3GPP term used to identify the protocol layers and functionalityrelated to the core network. Typically, it includes the Call Control and Mobility Manage-ment layers. The corresponding 3GPP2 term is Upper Layers.

The radio interface of UMTS/IMT-2000, which is based on WCDMA, is layered intothree protocol layers [44]:

• the physical layer (L1),

• the data link layer (L2),

• network layer (L3).

Layer 1 comprises the WCDMA physical layer while layer 2 is composed of mediumaccess control (MAC), radio link control (RLC-C for the control plane and RLC-U for theuser plane), Broadcast/Multicast Control (BMC) and Packet Data Convergence (PDCP)protocols, as well as the link-access control (LAC) protocol. MAC and RLC belong to theaccess stratum and terminate within UTRAN whereas it is proposed that LAC belongs tothe nonaccess stratum and terminates in the core network. The network layer of the controlplane is split into the radio resource control (RRC) sublayer and the mobility management(MM) and connection management (CM) sublayers. CM and MM belong to the nonaccessstratum while RRC belongs to the access stratum.

Like RLC, Layer 3 is also divided into Control (C-) and User (U-) planes. In the C-plane,Layer 3 is partitioned into sublayers where the lowest sublayer, denoted as RadioResource Control (RRC), interfaces with layer 2 and terminates in the UTRAN. The nextsublayer, denoted ’Duplication avoidance’, terminates in the core network, is part of theAccess Stratum; it provides the Access Stratum Services to higher layers.

The functions and services of each protocol layer will be exemplified in the following sec-tion [45]:

6.2.1 Physical Layer

Physical layer offers information transfer services to the MAC layer. These services aredenoted as transport channels, which will be mentioned later.

The physical layer comprises at least the following functions:

• forward error-correction coding, interleaving, and rate matching;

• measurements and indication to higher layer (e.g. FER, SIR, interference power, trans-mission power, etc ...);

• error detection on transport channels;

• macrodiversity distribution/combining and soft handover execution;

• multiplexing of transport channels and demultiplexing of coded composite transportchannels;

• mapping of coded composite transport channels;


• modulation and spreading/demodulation, and despreading of physical channels;

• frequency and time (chip, bit, slot, and frame) synchronization;

• closed-loop power control;

• power weighting and combining of physical channels;

• radio frequency (RF) processing.

6.2.2 MAC Layer

The MAC layer offers data transfer service to RLC and higher layers and it comprises atleast the following functions:

• selection of appropriate transport format (TF, basically bit rate), within a predefinedset, per information unit delivered to the physical layer;

• service multiplexing on RACH, FACH, and dedicated channels;

• priority handling between data flows of one user as well as between data flows fromseveral users-the latter being achieved by means of dynamic scheduling;

• access control on RACH;

• address control on RACH and FACH;

• contention resolution on RACH.

6.2.3 RLC Layer

The RLC layer offers the following services to the higher layers:

• layer 2 connection establishment/release;

• transparent data transfer, i.e., no protocol overhead is appended to the information unitreceived from the higher layer;

• assured and unassured data transfer.

The RLC layer comprises at least the following functions:

• segmentation and assembly;

• transfer of user data;

• error correction by means of retransmission optimized for the WCDMA physical layer;

• sequence integrity (used by at least the control plane);

• duplicate detection;

• flow control.

6.2.4 RRC Layer

The RRC layer offers the core network the following services:


• general control service, which is used as an information broadcast service;

• notification service, which is used for paging and notification of a selected MS(’s);

• dedicated control service, which is used for establishment/release of a connection andtransfer of messages using the connection.

The RRC layer comprises at least the following functions:

• broadcasting of system information;

• radio resource handling (e.g., code allocation, handover, admission control, and mea-surement reporting/control);

• control of requested quality of service;

6.3 Types and Structures of Channels

There are three types of channels defined in WCDMA for FDD mode: transport, logicaland physical channels.

6.3.1 Transport Channels

Transport channels are defined by how they are transmitted over the radio interface. Theyare the services mapped onto physical channels and offered by Layer 1 to the higher lay-ers. Each transport channel has a set of characteristics and transports logical channels. Thefollowing is a list of the characteristics of the information to be sent over the radio inter-face:

• Format: Encoding (convolution, block or turbo codes), Interleaving, Bit rate

• Framing / Multiplexing: How the information is multiplexed if it is composed of sev-eral sources.

• General Characteristics:

- Uplink or Downlink

- Power control characteristics

- Risk of collision or not

- Mobile Station identification method (in-band or inherent)

- Possibility of beam forming

- Data rate variations

- Broadcast area (entire cell or selected sector only)

A general classification of transport channels is into two groups:

- Dedicated transport channels (where the MSs are identified by the physical channel, i.e.code and frequency for FDD and code, time slot and frequency for TDD). There existsonly one type of dedicated transport channel, the Dedicated Channel (DCH).


• The Dedicated Channel (DCH) - is a downlink or uplink transport channel. The DCH istransmitted over the entire cell or over only a part of the cell using beam-forming anten-nas. The Dedicated Channel (DCH) is characterized by the possibility of fast ratechange (every 10ms), fast power control and inherent addressing of MSs.

- Common transport channels (where there is a need for inband identification of the MSswhen particular MSs are addressed). There are six types of common transport channels:BCH, FACH, PCH, RACH, CPCH and DSCH.

• The Broadcast Channel (BCH) - is a downlink transport channel that is used to broad-cast system and cell specific information. The BCH is always transmitted over theentire cell with a low fixed bit rate.

• The Forward Access Channel (FACH) - is a downlink transport channel. The FACH istransmitted over the entire cell or over only a part of the cell using beam-forming anten-nas. It uses slow power control and designed to carry control information specific to amobile station when the network knows in what cell the mobile is located.

• The Paging Channel (PCH) - is a downlink transport channel that carries control infor-mation specific to a mobile station when the network does not know where the mobilestation is located. In that case, the paging message is broadcasted over the entire net-work. A response from the mobile indicates its location cell. The transmission of thePCH is associated with the transmission of a physical layer signal, the Paging Indicator,to support efficient sleep-mode procedures.

• The Random Access Channel (RACH) - is a contention based uplink channel used fortransmission of relatively small amount of data, e.g. for initial access or non-realtimededicated control or traffic data. The RACH is always received from the entire cell,characterized by a limited size data field, a collision risk and by the use of open looppower control.

• The Common Packet Channel (CPCH) - is an uplink transport channel and only existsin FDD mode. The CPCH is a contention based random access channel used for trans-mission of bursty data traffic. CPCH is associated with a dedicated channel on thedownlink which provides power control for the uplink CPCH. The CPCH is fast powercontrolled.

• The Downlink Shared Channel (DSCH) - is a downlink transport channel shared byseveral MSs carrying dedicated control or traffic data. The DSCH is associated with aDCH.

6.3.2 Logical Channels

An information stream dedicated to the transfer of a specific type of information over theradio interface is called a Logical Channel. The Layer2/MAC sublayer provides datatransfer services to higher layers on logical channels which are mapped onto transportchannels. A set of logical channel types is defined for different kinds of data transfer ser-vices as offered by MAC and classified into:

- Control channels (for transfer of control information)


- Traffic channels (for transfer of user plane information)

The control channels are:

• Broadcast Control Channel (BBCH - DL): Broadcasts control information from basestation (BS) to mobile stations (MS).

• Paging Control Channel (PCCH - DL): Transfers paging information from BS to MS.

• Common Control Channels (CCCH - DL & UL)

- Forward Access Channel (FACH - DL): Carries control information from BS to MSwhen the network knows where the MS is located on the network.

- Random Access Channel (RACH - UL): Channel that carries control informationfrom the MS to the BS. (contention channel)

• Dedicated Control Channel (DCCH - DL & UL): Point-to-point bi-directional channelthat transfers dedicated control information from a BS to a MS or vice versa.

The Traffic channel is:

• Dedicated Traffic Channel (DTCH - DL & UL): Point-to-point bi-directional channelthat carries user data.

The mappings between logical channels and transport channels as seen from the MS andUTRAN sides are shown in figure 16 and figure 17 respectively.

FIGURE 16. Logical channels mapped onto transport channels, seen from the MS side.


FIGURE 17. Logical channels mapped onto transport channels seen from the UTRAN side.

6.3.3 Physical Channels

Physical channels are defined differently for FDD and TDD. For FDD, a physical channelis defined by its carrier frequency, access code and in the uplink, by the relative phase ofthe signal (either the In-Phase or Quadrature component). Similarly, TDD defines a phys-ical channel by its carrier frequency, access code, relative phase for the uplink and also bythe time slot in which it is transmitted. Physical channels are used to carry the transportchannels through the air interface.

Being more widely used, FDD is the mode that is mainly discussed in this report. Itrequires the allocation of two frequency bands: one for the uplink and another for thedownlink. It has the advantage of being able to transmit and receive at the same time. Fur-thermore, the size of the cell is not limited by propagation delays like in TDD because ofthe absence of time slots and guard periods, which also makes the timing sychronisationbetween base and mobiles less critical than TDD. Because it transmits and receive at thesame time, FDD radio units need duplexers in order to separate the incoming and outgoingsignals at the antenna. Duplexers are made of filters which increase the complexity andcost of the hardware. Moreover, FDD does not allocate efficiently the available bandwidthfor all types of services. For example, Internet access requires more throughput on thedownlink than on the uplink. Of course by adjusting the spreading factor, it becomes pos-sible to use only the required data rate, but it is still impossible to trade uplink bandwidthfor downlink bandwidth.

A general classification of physical channels is into two groups: dedicated and commonphysical channels. Two types of dedicated physical channels exist:

• Dedicated physical data channel (DPDCH - UL/DL) : Carries data generated at Layer 2and above.

• Dedicated physical control channel (DPCCH - UL/DL) : Carries data generated atLayer 1 (pilot bits, TPC commands, and optional transport-format information)

Each connection is allocated one DPCCH and zero, one , or several DPDCH’s.

Common physical channels that are classified for downlink are:


• Primary common control channel (primary CCPCH), DL - Fixed rate (32kbps,SF=256) channel, designed to carry the BCCH.

• Secondary common control channel (secondary CCPCH), DL - Carries the FACH andthe PCH.

• Synchronization channel (SCH), DL - provides timing information and is used for han-dover measurements by the mobile station.

• Common Pilot Channel (CPICH), DL - is a fixed rate (30 kbps, SF=256) downlinkphysical channel that carries a pre-defined bit/symbol sequence.

• Physical Downlink Shared Channel (PDSCH), DL - used to carry the Downlink SharedChannel (DSCH), is shared by users based on code multiplexing. As the DSCH isalways associated with a DCH, the PDSCH is always associated with a downlinkDPCH.

• Acquisition Indication Channel (AICH), DL - is used to carry Acquisition Indicators(AI).

• Page Indication Channel (PICH), DL - is a fixed rate (SF=256) physical channel usedto carry the Page Indicators (PI). The PICH is always associated with an S-CCPCH towhich a PCH transport channel is mapped.

Common physical channels that are classified for uplink are:

• Physical random access channel (PRACH), UL - carries the RACH.

• Physical Common Packet Channel (PCPCH), UL - is used to carry the CPCH.

The mappings between physical channels and transport channels are shown in the tablebelow.

TABLE 3. Transport Channel to Physical Channel Mapping

Transport Channels Physical Channels

BCH Common Pilot Channel (CPICH)

Primary Common Control Physical Channel (P-CCPCH)

FACH and PCH Secondary Common Control Physical Channel (S-CCPCH)

RACH Physical Random Access Channel (PRACH)

CPCH Physical Common Packet Channel (PCPCH)

DCH Dedicated Physical Data Channel (DPDCH)

Dedicated Physical Control Channel (DPCCH)

Synchronisation Channel (SCH)

DSCH Physical Downlink Shared Channel (PDSCH)

Page Indication Channel (PICH)

Acquisition Indication Channel (AICH)


The DCHs are coded and multiplexed, and the resulting data stream is mapped sequen-tially (first-in-first-mapped) directly to the physical channel(s). The mapping of BCH andFACH/PCH is equally straightforward, where the data stream after coding and interleav-ing is mapped sequentially to the Primary and Secondary CCPCH respectively. Also forthe RACH, the coded and interleaved bits are sequentially mapped to the physical chan-nel, in this case the message part of the random access burst on the PRACH.

6.4 Physical Channels in Detail

6.4.1 Uplink Physical Channels

There are two dedicated channels and two common channels on the uplink. User data istransmitted on the dedicated physical data channel (DPDCH), and control information istransmitted on the dedicated physical control channel (DPCCH).

The Random Access Channel (RACH) is a common access channel. and is based on aSlotted ALOHA approach with fast acquisition indication. The MS can start the transmis-sion at a number of well-defined time-offsets, denoted access slots. There are 15 accessslots per two frames and they are spaced 5120 chips apart. Timing information on theaccess slots and the acquisition indication can be found in [46]. Figure 18 shows theaccess slot numbers and their spacing to each other. Information on what access slots areavailable in the current cell is given by higher layers.

FIGURE 18. RACH access slot numbers and their spacing

Before the transmission of a random access request, the mobile terminal should carry outthe following tasks:

• Achieve chip, slot, and frame synchronization to the target base station from the syn-chronization channel (SCH) and obtain information about the downlink scramblingcode also from the SCH

• Retrieve information from BCCH about the random access code(s) used in the targetcell/sector


• Estimate the downlink path loss, which is used together with a signal strength target tocalculate the required transmit power of the random access request

It is possible to transmit a short packet together with a random access burst without sett-ting up a scheduled packet channel. No separate access channel is used for packet trafficrelated random access, but all traffic shares the same random access channel. More thanone random access channel can be used if the random access capacity requires such anarrangement. The performance of the selected solution is presented in [47].

The Common Packet Channel (CPCH) transmission is based on DSMA-CD (DigitalSense Multiple Access with Collision Detection) approach with fast acquisition indica-tion. The MS can start transmission at a number of well-defined time-offsets, relative tothe frame boundary of the received BCH of the current cell. The access slot timing andstructure is identical to RACH. The structure of the CPCH random access transmission isshown in figure 19. The CPCH random access transmission consists of one or severalAccess Preambles [A-P] of length 4096 chips, one Collision Detection Preamble (CD-P)of length 4096 chips, a [10] ms DPCCH Power Control Preamble (PC-P) and a message ofvariable length Nx10 ms.

FIGURE 19. Structure of CPCH random access transmission

The principle frame structure of the uplink Dedicated Physical Data Channel (DPDCH)shown in figure 20 [2].

FIGURE 20. WCDMA uplink multirate transmission.


Each DPDCH frame on a single code carries 160 x 2k bits (16 x 2k Kb/s), where k = 0,1,

..., 6, corresponding to a spreading factor of 256/2k with the 4.096-Mc/s chip rate. Multi-ple parallel variable rate services (= dedicated logical traffic and control channels) can betime multiplexed within each DPDCH frame. The overall DPDCH bit rate is variable on aframe-by-frame basis.

In most cases, only one DPDCH is allocated per connection, and services are jointly inter-leaved sharing the same DPDCH. However, multiple DPDCHs can also be allocated (e.g.to avoid a too low spreading factor at high data rates).

The Dedicated Physical Control Channel (DPCCH) is needed to transmit pilot symbolsfor coherent reception, power control signaling bits, and rate information (transport formatindicator, TFI) for rate detection. A certain transport format (TF) defines how the layer 2data carried on the DPDCH(‘s) is multiplexed and coded and what spreading factor is used

etc. Two basic solutions for multiplexing physical control and data channels are time mul-

tiplexing and code multiplexing. A combined IQ and code multiplexing solution (dual-

channel QPSK) is used in WCDMA uplink to avoid electromagnetic compatibility (EMC)

problems with discontinuous transmission (DTX).

The major drawback of the time multiplexed control channel are the EMC problems that

arise when DTX is used for user data. One example of a DTX service is speech. During

silent periods no information bits need to be transmitted, which results in pulsed transmis-

sion as control data must be transmitted in any case. This is illustrated in figure 21 [2].

Because the rate of transmission of pilot and power control symbols is on the order of 1 to

2 kHz, they cause severe EMC problems to both external equipment and terminal interi-

ors. This EMC problem is more difficult in the uplink direction since mobile stations can

be close to other electrical equipment, like hearing aids.

FIGURE 21. Illustration of pulsed transmission with multiplexed control channel.

The IQ/code multiplexed control channel is shown in figure 22 [2]. Now, since pilot and

power control are on a separate channel, no pulse-like transmission takes place. Interfer-

ence to other users and cellular capacity remains the same as in the time multiplexed solu-

tion. In addition, link-level performance is the same in both schemes if the energy

allocated to the pilot and the power control bits is the same.


FIGURE 22. Illustration of parallel transmission of DPDCH and DPCCH.

6.4.2 Downlink Physical Channels

In the downlink, there are seven common physical channels. The dedicated channels(DPDCH and DPCCH) are time multiplexed. The EMC problem caused by discontinuoustransmission is not considered difficult in downlink since (1) there are signals to severalusers transmitted in parallel and at the same time and (2) base stations are not so close toother electrical equipment, like hearing aids.

In the downlink, time multiplexed pilot symbols are used for coherent detection. Since thepilot symbols are connection dedicated, they can be used for channel estimation withadaptive antennas as well. Furthermore, the connection dedicated pilot symbols can beused to support downlink fast power control.

Figure 23 shows the frame structure of the downlink DPCH. Each frame of length 10 ms issplit into 15 slots, each of length Tslot = 2560 chips, corresponding to one power-controlperiod. A super frame corresponds to 72 consecutive frames, i.e. the super-frame length is720 ms.

FIGURE 23. Frame structure of downlink DPCH.

The Primary Common Control Physical Channel (P-CCPCH) carries the BCCH chan-nel and a time multiplexed common pilot channel which can be used for coherent detec-tion. It is of fixed rate and is mapped to the DPDCH in the same way as dedicated trafficchannels. The primary CCPCH is allocated the same channelization code in all cells. Amobile terminal can thus always find the BCCH, once the base station’s unique scrambling


code has been detected during the initial cell search. Figure 24 shows the frame structureof the Primary CCPCH. The frame structure differs from the downlink DPCH in that noTPC commands, no TFCI and no pilot bits are transmitted The Primary CCPCH is nottransmitted during the first 256 chips of each slot. Instead, Primary SCH and SecondarySCH are transmitted during this period.

FIGURE 24. Frame structure for Primary Common Control Physical Channel.

The Secondary Common Control Physical Channel (S-CCPCH) carries the PCH andFACH in time multiplex within the super frame structure. The rate of the S-CCPCH maybe different for different cells and is set to provide the required capacity for PCH andFACH in each specific environment. The channelization code of the secondary CCPCH istransmitted on the primary CCPCH. There are two types of Secondary CCPCH: those thatinclude TFCI and those that do not include TFCI. It is the UTRAN that determines if aTFCI should be transmitted, hence making it mandatory for all MSs to support the use ofTFCI. The frame structure of the Secondary CCPCH is shown in figure 25.

FIGURE 25. Frame structure for Secondary Common Control Physical Channel


The parameter k in figure 25 determines the total number of bits per downlink SecondaryCCPCH slot. It is related to the spreading factor SF of the physical channel as SF = 256/

2k. The spreading factor range is from 256 down to 4.

The values for the number of bits per field are given in table 4 and table 5, which are takenfrom [46]. The channel bit and symbol rates given in Table 16 are the rates immediatelybefore spreading. The pilot patterns are also given in [46] as well.

The FACH and PCH can be mapped to the same or to separate S-CCPCHs. If FACH andPCH are mapped to the same S-CCPCH, they can be mapped to the same frame. The maindifference between a CCPCH and a downlink dedicated physical channel is that a CCPCHis not inner-loop power controlled. The main difference between the Primary and Second-ary CCPCH is that the P-CCPCH has a fixed predefined rate while the S-CCPCH can sup-port variable rate with the help of the TFCI field included. Furthermore, a P-CCPCH iscontinuously transmitted over the entire cell while a S-CCPCH is only transmitted whenthere is data available and may be transmitted in a narrow lobe in the same way as a dedi-cated physical channel (only valid for a S-CCPCH carrying the FACH).

* If TFCI bits are not used, then DTX shall be used in TFCI field.

TABLE 4. Secondary CCPCH fields with pilot bits

Slot format

#i

Channel Bit

Rate (kbps)

Channel

Symbol Rate

(ksps)

SF Bits/

Frame

Bits/

Slot

Ndata Npilot NTFCI

0 30 15 256 300 20 12 8 0

1 30 15 256 300 20 10 8 2

2 60 30 128 600 40 32 8 0

3 60 30 128 600 40 30 8 2

4 120 60 64 1200 80 64 8 8*

5 240 120 32 2400 160 144 8 8*

6 480 240 16 4800 320 296 16 8*

7 960 480 8 9600 640 616 16 8*

8 1920 960 4 19200 1280 1256 16 8*


* If TFCI bits are not used, then DTX shall be used in TFCI field.

The Synchronization Channel (SCH) consists of two subchannels, the primary and sec-ondary SCHs. Figure 26 illustrates the structure of the SCH radio frame. The SCH appliesshort code masking to minimize the acquisition time of the long code [48]. The SCH ismasked with two short codes (primary and secondary SCH). The unmodulated primarySCH is used to acquire the timing for the secondary SCH. The modulated secondary SCHcode carries information about the long code group to which the long code of the BSbelongs. In this way, the search of long codes can be limited to a subset of all the codes.

FIGURE 26. Structure of the synchronization channel (SCH)

The primary SCH consists of an unmodulated code of length 256 chips, which is transmit-ted once every slot. The primary synchronization code is the same for every base station inthe system and is transmitted time aligned with the slot boundary, as illustrated in figure26 [2].

TABLE 5. Secondary CCPCH fields without pilot bits

Slot format

#i

Channel Bit

Rate (kbps)

Channel

Symbol Rate

(ksps)

SF Bits/

Frame

Bits/

Slot

Ndata Npilot NTFCI

0 30 15 256 300 20 20 0 0

1 30 15 256 300 20 18 0 2

2 60 30 128 600 40 40 0 0

3 60 30 128 600 40 38 0 2

4 120 60 64 1200 80 72 0 8*

5 240 120 32 2400 160 152 0 8*

6 480 240 16 4800 320 312 0 8*

7 960 480 8 9600 640 632 0 8*

8 1920 960 4 19200 1280 1272 0 8*


The secondary SCH consists of one modulated code of length 256 chips, which is trans-mitted in parallel with the primary SCH. The secondary synchronization code is chosenfrom a set of 16 different codes depending on to which of the 32 different code groups thebase station downlink scrambling code belongs.

The secondary SCH is modulated with a binary sequence of length 16 bits, which isrepeated for each frame. The modulation sequence, which is the same for all base stations,has good cyclic autocorrelation properties.

The SCH is nonorthogonal to the other downlink physical channels.

The frame structure of the Common Pilot Channel (CPICH) is shown in figure 27.There are two types of Common pilot channels, the Primary and Secondary CPICH. Theydiffer in their use and the limitations placed on their physical features.

FIGURE 27. Frame structure for Common Pilot Channel.

The Primary Common Pilot Channel has the following characteristics [46]:

• The same channelization code is always used for this channel

• Scrambled by the primary scrambling code

• One per cell

• Broadcast over the entire cell

The Primary CPICH is the phase reference for the following downlink channels: SCH,Primary CCPCH, AICH, PICH. The Primary CPICH is also the default phase referencefor all other downlink physical channels.

A Secondary Common Pilot Channel characteristics [46] :

• Can use an arbitrary channelization code of SF=256

• Scrambled by either the primary or a secondary scrambling code


• Zero, one, or several per cell

• May be transmitted over only a part of the cell

• A Secondary CPICH may be the reference for the Secondary CPCCH and the downlink DPCH. If this is the case, the MS is informed about this by higher-layer signalling.

The Physical Downlink Shared Channel (PDSCH), is used to carry the DownlinkShared Channel (DSCH), is shared by users based on code multiplexing. As the DSCH isalways associated with a DCH, the PDSCH is always associated with a downlink DPCH.The frame and slot structure of the PDSCH are shown on figure 28.

FIGURE 28. Frame structure for the PDSCH.

To indicate for MS that there is data to decode on the DSCH, two signalling methods arepossible, either using the TFCI field, or higher layer signalling.

The PDSCH transmission with associated DPCH is a special case of multicode transmis-sion. The PDSCH and DPCH do not have necessary the same spreading factors and forPDSCH the spreading factor may vary from frame to frame. The relevant Layer 1 controlinformation is transmitted on the DPCCH part of the associated DPCH, the PDSCH doesnot contain physical layer information. The channel bit and symbol rates for PDSCH aregiven in table 6.


For PDSCH the allowed spreading factors may vary from 256 to 4.

If the spreading factor and other physical layer parameters can vary on a frame-by-framebasis, the TFCI shall be used to inform the MS what are the instantaneous parameters ofPDSCH including the channelisation code from the PDSCH OVSF (Orthhogonal VariableSpreading Factor) code tree.

A DSCH may be mapped to multiple parallel PDSCHs as well, as negotiated at higherlayer prior to starting data transmission. In such a case the parallel PDSCHs shall be oper-ated with frame synchronization between each other.

The Acquisition Indicator channel (AICH) is used to carry Acquisition Indicators (AI).Acquisition Indicator AIi corresponds to signature i on the PRACH or PCPCH. Note thatfor PCPCH, the AICH is either in response to an access preamble or a CD preamble. Thecorresponding to the access preamble AICH is the AP-AICH and the corresponding to theCD preamble AICH is the CD-AICH. The AP-AICH and CD-AICH use different channel-ization codes.

FIGURE 29. Structure of Acquisition Indicator Channel (AICH).

Figure 29 illustrates the frame structure of the AICH. Two AICH frames of total length 20ms consist of 15 access slots (AS), each of length 20 symbols (5120 chips). Each access

TABLE 6. PDSCH fields

Slot format

#i

Channel Bit

Rate (kbps)

Channel

Symbol Rate

(ksps)

SF Bits/

Frame

Bits/

Slot

Ndata

0 30 15 256 300 20 20

1 60 30 128 600 40 40

2 120 60 64 1200 80 80

3 240 120 32 2400 160 160

4 480 240 16 4800 320 320

5 960 480 8 9600 640 640

6 1920 960 4 19200 1280 1280


slot consists of two parts, an Acquisition-Indicator (AI) part and an empty part. The emptypart of the access slot consists of 4 zeros. The phase reference for the AICH is the CPICH.

The frame structure of the Page Indicater Channel (PICH) is illustrated in Figure 30.One PICH frame of length 10 ms consists 300 bits. Of these, 288 bits are used to carryPage Indicators. The remaining 12 bits are not used.

FIGURE 30. Structure of Page Indicator Channel (PICH)

6.5 Spreading

The WCDMA scheme employs long spreading codes. Different spreading codes are usedfor cell separation in the downlink and user separation in the uplink. In the downlink, Goldcodes of length 218 are used, but they are truncated to form a cycle of a 10-ms frame. Thetotal number of available scrambling codes is 512, divided into 32 code groups with 16codes in each group to facilitate a fast cell search procedure. In the uplink, either short orlong spreading (scrambling codes) are used. The short codes are used to ease the imple-mentation of advanced multiuser receiver techniques; otherwise long spreading codes canbe used. Short codes are VL-Kasami codes of length 256 and lond codes are Goldsequences of length 241, but the latter are truncated to form a cycle of a 10-ms frame.

FIGURE 31. IQ/code multiplexing with complex spreading circuit.

For channelization, orthogonal codes are used. Orthogonality between the differentspreading factors can be achieved by the tree-structured orthogonal codes.

IQ/code multiplexing leads to parallel transmission of two channels, and therefore, atten-tion must be paid to modulated signal constellation and related peak-to-average powerratio (crest factor). By using the complex spreading circuit shown in figure 31 [2], the


transmitter power amplifier efficiency remains the same as for QPSK transmission in gen-eral.

Moreover, the efficiency remains constant irrespective of the power difference G betweenDPDCH and DPCCH. This can be explained with figure 32 [2], which shows the signalconstellation for IQ/code multiplexed control channel with complex spreading. In the mid-dle constellation with G = 0.5 all eight constellation points are at the same distance fromthe origin. The same is true for all values of G. Thus, signal envelope variations are verysimilar to the QPSK transmission for all values of G. The IQ/code multiplexing solutionwith complex scrambling results in power amplifier output backoff requirements thatremain constant as a function of power difference. Furthermore, the achieved output back-off is the same as for one QPSK signal.

FIGURE 32. Signal constellation for IQ/code multiplexed control channel with complex spreading. G is the power difference between DPCCH and DPDCH.

6.6 Channel Coding and Multiplexing

A key feature of the WCDMA radio interface is the possibility to transport multiple paral-lel services (TrCH’s) with different quality requirements on one connection. The basic

scheme for the channel coding and transport channel multiplexing in WCDMA is drawn

in figure 33 [45] Parallel TrCH’s (TrCH-1 to TrCH-M) are seperately channel coded and

interleaved. The coded TrCH’s are then time multiplexed into a coded composite TrCH

(CC-TrCh). Final intraframe (10 ms) interleaving is carried out after transport channel

multiplexing. After service multiplexing and channel coding, the multiservice data stream

is mapped to one DPDCH. If the total rate exceeds the upper limit for single code trans-

mission, several DPDCHs can be allocated.


FIGURE 33. Channel coding and transport-channel multiplexing.

A second alternative for service multiplexing would be to map parallel services to differ-ent DPDCHs in a multicode fashion with separate channel coding/interleaving. With thisalternative scheme, the power, and consequently the quality of each service, can be sepa-rately and independently controlled. The disadvantage is the need for multicode transmis-sion, which will have an impact on mobile station complexity due to the increasedenvelope variations in the transmitted signal and the need for multiple RAKE receivers.Multicode transmission sets higher requirements for the power amplifier linearity in trans-mission, and more correlators are needed in reception.

6.6.1 Channel Coding

Different coding and interleaving schemes can be applied to a TrCH depending on thespecific requirements in terms of error rates, delay etc. The following channel codingschemes are used:

• Rate 1/2 convolutional coding is typically applied for low-delay services (like BCH,PCH, FACH and RACH transport channels) with moderate error rate requirements

(BER = 10-3).

• A concatenation of rate 1/3 convolutional coding and outer Reed-Solomon coding +interleaving can be applied for high-quality services (like CPCH and DCH with BER

between 10-3 and 10-6). Retransmissions can be utilized to guarantee service quality fornon real-time packet data services.

• Turbo codes of rate 1/3 can also be applied for high rate high quality services (like

CPCH and DCH with BER between 10-3 and 10-6).


6.6.2 Rate Matching

After channel coding and service multiplexing, the total bit rate can be almost arbitrary.Rate matching adapts this rate to the limited set of possible bit rates of a DPDCH. Repeti-tion or puncturing is used to match the coded bit stream to the channel gross rate. Asshown in figure 33, there are two different rate-matching steps: static and dynamic ratematching.

. Static Rate Matching - This kind of matching is carried out at the addition, removal andredefinition of TrCH’s, i.e., on a very slow basis. The static rate matching is applied after

channel coding and uses code puncturing to adjust the channel coding rate of each TrCh

so that the maximum bit rate of the coded composite TrCh is matched to the bit rate of the

physical channel. In general, code puncturing is chosen for bit rates less than 20 percent

above the closest lower DPDCH bit rate. The static rate matching is applied both the

uplink and downlink. On the downlink, the static rate is used to, if possible, reduce the

coded composite TrCH rate to the closest lower physical channel rate (closest higher

spreading factor) thus avoiding the “overallocation” of orthogonal codes on the downlink

and reducing the risk for a code limited downlink capacity. The static rate matching

should be distributed between the parallel TrCH’s in such a way that the TrCH’s fulfill

their quality requirements at approximately the same channel signal-to-interference ratio

(SIR), i.e., the static rate matching also performs “SIR matching.”

. Dynamic Rate Matching - The dynamic rate matching is carried out once every 10 ms

radio frame, i.e., on a very fast basis. The dynamic rate matching is applied after transport

channel multiplexing and uses symbol repetition so that the instaneous bit rate of coded

composite TrCH is exactly matched to the bit rate of the physical channel. The dynamic

rate matching is only applied to the uplink. On the downlink, discontinious transmission

(DTX) within each slot is used when the instantaneous rate of the coded composite TrCh

does not exactly match the bit rate of the physical channel.

In short, for the uplink, rate matching is based on both symbol repetition and code punc-

turing, while it’s only based on code puncturing for the downlink since DTX is used oth-

erwise.

It should be noted that although the transport channel coding and multiplexing is carried

out by the physical layer, the process is fully controlled by the radio resource controller,

e.g, in terms of choosing the appropriate coding scheme, interleaving parameters, and rate

matching parameters.

6.7 Radio Resource Functions

6.7.1 Power Control

WCDMA employs fast closed-loop power control in both the uplink and downlink. The

basic power-control rate is 1600 Hz, and the power-control step can be varied adaptively

according to the MS speed and operating environment. SIR-based power control is used,

i.e., the receiver compares the estimated received SIR with a SIR target value and com-


mands the transmitter to increase or decrease the power accordingly. The power-controlcommand increases or decreases the power of all physical channels on one connection.

The target SIR values are controlled by an outer power control loop. This outer loop mea-sures the link quality, typically, a combination of frame and bit error rates (BER’s)

depending on the service, and adjusts the SIR targets accordingly. Ensuring that the lowest

possible SIR target is used at all times results in maximum capacity. In addition, the outer

loop is used to independently control the relative power of different physical channels

belonging to the same connection. As an example, the DPDCH and DPCCH power differ-

ence can be controlled by the outer loop to take into account the variations in DPDCH

coding gain for different environments.

Open-loop power control is used by the random-access procedure, where uplink path loss

is estimated from downlink path loss. Also, common-channel packet transmissions

depend on open-loop power control.

6.7.2 Random Access

A fast and efficient random-access scheme is essential for the UMTS system since packet

access is becoming more important in the third-generation systems. This will lead to an

increased number of random-access attempts that need to be served quickly.

The WCDMA random-access procedure is based on slotted ALOHA and works as fol-

lows.

• The MS achieves chip and frame synchronization to the target cell using the initial cell-search procedure.

• The BCCH is read to retrieve information about the random-access code(s) used in thetarget cell.

• The downlink path loss is estimated, and the estimate is used to calculate the requiredtransmit power of the random-access burst.

• A random-access burst is transmitted with a random time offset. The time offset is amultiple of 1.25 ms relative to the received frame boundary.

• The base station responds with an acknowledgment on the FACH.

• If the MS receives no acknowledgment, it selects a new time offset and tries again.

The random-access procedure is described in more detail in [49] and [50].

6.7.3 Initial Cell Search

WCDMA base stations are, in general, mutually asynchronous, i.e., there is no universaltime reference known to all base stations. To separate different cells, different downlinkscrambling codes are used. During the initial cell search, the MS first searches for thestrongest base-station cell. The MS then determines the scrambling code and the framesynchronization of that cell. The cell search consists of three steps.


In the first step, the MS acquires slot synchronization to the strongest base station. This isdone using a matched filter matched to the system-specific code used on the primary SCH.Each ray of each cell within hearable range will result in a peak in the signal output fromthe filter. The largest peak indicates the slot timing of the strongest cell.

In the second step, the MS correlates the received slot synchronized signal with the avail-able 16 codes used on the secondary SCH, followed by correlation with the 16 differentcyclic shifts of the 16-b modulation sequence. The maximum of these 256 correlation val-ues identifies the code group and the frame timing.

The third cell-search step consists of an exhaustive search of all the scrambling codes inthe code group identified in the second step. The search is done through symbol by symbolcorrelation over the primary CCPCH. Since frame synchronization was obtained in thesecond step, the start of the scrambling code is known.

When the scrambling code has been identified, the cell and system specific broadcastinformation on the primary CCPCH can be read.

6.7.4 Handover

The normal handover in WCDMA is soft intrafrequency handover, where the MS is con-nected to two or more cells simultaneously on the same frequency. The MS continuouslysearches for new cells, using the cell-search technique described above, but the search islimited to a list of neighboring cells broadcast from the network. The neighboring list tellsthe MS in which order to search for the scrambling codes, and it can also limit the searchto a subset of all available codes. In soft handover, the uplink signals are combined in thenetwork, and downlink combining of signals is done in the MS’s RAKE receiver.

When including a new additional base station in the active set, i.e., the set of base stations

currently connected, the MS signals via the old link(s), how the new base station should

adjust its DPCCH/DPDCH frame timing to minimize the received frame timing differ-

ences in the MS. This can be done since the MS from the cell search knows the relative

frame timing of the primary CCPCH of the handover candidates. Timing adjustments of

dedicated downlink channels (DPCCH/DPDCH) of the new base station relative to the

primary CCPCH can be carried out with a resolution of one DPCCH/DPDCH symbol

without losing orthogonality of downlink codes. The synchronization of the dedicated

downlink signals from the two base stations with an accuracy of one symbol, enables the

mobile RAKE receiver to collect the macro diversity energy from the two base stations.

A special case of soft handover is the softer handover, where the MS is connected to two

cells belonging to one base-station site. Instead of doing the uplink combining in the net-

work, as is the case for soft handover, softer handover combining can be done in the base

station. This makes it possible to use more efficient uplink combining, e.g., maximum

ratio combining.

In WCDMA, soft and softer handover use relative handover thresholds. By doing so,

fewer MS’s will be in soft or softer handover, compared to when absolute thresholds are


used, as is the case for current narrow-band CDMA systems, i.e., IS-95-A. Moreover, theadding and dropping of cells in the active set is load dependent for IS-95-A, while theactive set updating in WCDMA is load independent. This behavior is illustrated in figure34 [45].

FIGURE 34. Comparison of active set updating for absolute (IS-95-A) and relative (WCDMA) handover thresholds. A and B are candidate cells.

Furthermore, as softer handover can employ more efficient combining in the uplink andhas lower network transmission load, the handover margin for softer handover will typi-cally be larger than for soft handover. The handover parameters are service and loaddependent. Even though much of the handover functionality resides in the MS, the net-work can still put a veto on the MS’s suggestion of cells to connect to.

As mentioned, the normal handover in WCDMA is a soft intrafrequency handover. How-

ever, interfrequency handovers are also supported. Interfrequency handovers in the sys-

tem are essential to support:

• hot-spot scenarios, where a cell uses more carriers than the surrounding cells;

• hierarchical cell structures, where macro, micro, and pico layers are on different fre-quencies;

• handovers between different operators;

• handovers to other systems, e.g., GSM.

To support seamless interfrequency handovers, measurements on other frequencies mustbe possible without disturbing the normal data flow. Since the MS is receiving the down-link signal continuously, there is no time to carry out measurements on other frequenciesusing the ordinary receiver. Two methods are considered for interfrequency measurementsin WCDMA:

• Dual receiver

• Slotted mode


The dual receiver approach is considered suitable especially if the mobile terminalemploys antenna diversity. During the interfrequency measurements, one receiver branchis switched to another frequency for measurements, while the other keeps receiving fromthe current frequency. The loss of diversity gain during measurements needs to be com-pensated for with higher downlink transmission power. The advantage of the dual receiverapproach is that there is no break in the current frequency connection. Fast closed looppower loop is running all the time.

The slotted mode is considered attractive for the mobile station without antenna diversity.The information normally transmitted during a 10-ms frame is compressed time either bycode puncturing or by changing the FEC rate. A 10-ms data frame can then be transmittedin less than 10 ms. The transmission is done with higher power than normal to compensatefor the decreased processing gain. Using this technique, an idle period of up to 5 ms is cre-ated during which no data is to be received by the MS. This period can then be used totune the receiver to other frequencies and signaal strength measurements on those.

Base stations in WCDMA need not be synchronized, and therefore, no external source ofsynchronization, like GPS, is needed for the base stations. Asynchronous base stationsmust be considered when designing soft handover algorithms and when implementingposition location services.

6.7.5 Interoperatibility Between GSM and WCDMA

The handover between the WCDMA system and the GSM system, offering worldwidecoverage already today, has been one of the main design criteria taken into account in theWCDMA frame timing definition. The GSM compatible multiframe structure, with asuperframe multiple of 120 ms, allows similar timing for intersystem measurements as inthe GSM system itself. Apparently the needed measurement interval does not need to beas frequent as for GSM terminal operating in a GSM system, as intersystem handover isless critical from intra-system interference point of view. Rather, the compatibility in tim-ing is important that when operating in WCDMA mode, a multimode terminal is able tocatch the desired information from the synchronization bursts in the synchronizationframe on a GSM carrier with the aid of frequency correction burst. This way the relativetiming between a GSM and WCDMA carriers is maintained similar to the timing betweentwo asynchronous GSM carriers.

FIGURE 35. Measurement timing relation between WCDMA and GSM frame structure.


The timing relation between WCDMA channels and GSM channels is indicated in figure35 [2], where the GSM traffic channel and WCDMA channels use similar 120 ms multi-frame structure.

The GSM frequency correction channel (FCCH) and GSM synchronization channel(SCH) use one slot out of the eight GSM slots in the indicated frames with the FCCHframe with one time slot for FCCH always preceding the SCH frame with one time slot forSCH as indicated in the figure 35.

A WCDMA terminal can do the measurements either by requesting the measurementintervals in a form of slotted mode where there are breaks in the downlink transmission orthen it can perform the measurements independently with a suitable measurement pattern.With independent measurements the dual receiver approach is used instead of the slottedmode since the GSM receiver branch can operate independently of the WCDMA receiverbranch.

For smooth interoperation between the systems, information needs to be exchangedbetween the systems, in order to allow WCDMA base station to notify the terminal of theexisting GSM frequencies in the area. In addition, more integrated operation is needed forthe actual handover where the current service is maintained, taking naturally into accountthe lower data rate capabilities in GSM when compared to UMTS maximum data ratesreaching all the way to 2 Mb/s.

The GSM system is likewise expected to be able to indicate also the WCDMA spreadingcodes in the area to make the cell identification simpler and after that the existing mea-surement practises in GSM can be used for measuring the WCDMA when operating inGSM mode.

As the WCDMA does not rely on any superframe structure as with GSM to find out syn-chronization, the terminal operating in GSM mode is able to obtain the WCDMA framesynchronization once the WCDMA base station scrambling code timing is acquired. Thebase station scrambling code has 10-ms period and its frame timing is synchronized toWCDMA common channels.

6.8 Medium Access Control and Radio Link Control

The MAC and RLC protocols are responsible for efficiently transferring data of both real-time and nonreal-time services. The transfer of nonreal-time data transfer includes thepossibility of an optimized low-level automatic repeat request (ARQ) at the RLC layer,offering higher protocol layers reliable data transfer. In addition, the MAC layer controlsthe multiplexing of data streams originating from different services. A description of theMAC/RLC protocol(s) can also be found in [51].

6.8.1 Data Flow

In order to achieve the requirements mentioned above, the RLC layer segments the datastreams into small packets, RLC protocol data units (RLC PDU’s) suitable for transmis-


sion over the radio interface. In figure 36 [45], the data flow of the WCDMA system isshown.

FIGURE 36. Segmentation and transformation of network layer protocol data units (N-PDU’s).

Network layer PDU’s (N-PDU’s) are first segmented into smaller packets and trans-

formed into LAC PDU’s. The LAC overhead typically consists of a service access point

identifier, sequence number for a higher level ARQ, and other fields. The overhead is typ-

ically in the range of three octets. Then, the LAC PDU’s are segmented into small packets,

which are transformed into RLC PDU’s. The RLC PDU header typically contains a

sequence number. The sequence number is used for the optimized fast ARQ. The data

flow of the WCDMA system is very similar to the data flow of GPRS [52]. However, one

important difference is that in the GPRS system, an RLC PDU always consists of four

bursts, while the code rate may vary.

Hence, the number of information bits of the RLC PDU’s in the GPRS system can vary.

However, once a segment of the LAC PDU is transformed into an RLC PDU with a partic-

ular code, then, at a later time, this segment cannot be transformed into another RLC PDU

with a different code. Thus, in case of retransmissions, the same segment of the LAC PDU

will be retransmitted with the same code rate.

In the WCDMA system, on the other hand, all RLC PDU’s have the same size, regardless

of the transmission rate. This means that since the transmission rate may change every 10

ms, the number of RLC PDU’s transferred per 10 ms varies. This is illustrated in figure

36, where in the first transport frame two RLC PDU’s can be conveyed. The second trans-

port frame, which in time is equally long, can only convey one RLC PDU since the rate

has been changed.

6.8.2 Model of Operation

1) Packet Data Services: In this section, the model of operation when packets are transmit-

ted in the uplink is described.. For the downlink, packet transmission will be done in a

very similar way.


In the WCDMA system, packet data can be transmitted in three ways. First, if a layer 3packet is generated, the MS may choose to transmit it on the RACH. The data is simplyappended to the access burst. This method is illustrated in figure 37 [45].

FIGURE 37. Packet transmission on the RACH.

Typically, this method, which is called common channel packet transmission, is chosen ifthe MS has only a small amount of data to transmit. No reservation scheme is used in themethod, so the overhead necessary to transmit a packet is kept to a minimum. The MSdoes not need to get assigned a channel, thus, the access delay is kept small as well. Theother method is illustrated in figure 38 [45].

FIGURE 38. Packet transmission on a dedicated channel.

Here, the MS first sends a Resource Request (Res_Req) message. Typically, this is donewhen the packet is large. In this Res_Req message, an indication is given of what sort oftraffic is to be transmitted. The network then evaluates whether the MS can be assignedthe necessary resources. If that is the case, it transmits a Resource Allocation (Res_All)message on the FACH. A Res_All message consists of a set of transport formats (TF’s).

Out of this set, the MS will use a TF to transmit its data on the DCH. Exactly which TF the

MS may use and at what time the MS may initiate its transmission is either transmitted

together with the Res_All message or is indicated in a separate Capacity Allocation

(Cap_All) message at a later time. In situations where the traffic load is low, probably the

first alternative will be used, while the second alternative is used in cases where the load is

high and the MS is not allowed to immediately transmit. In figure 39 [45], the second

alternative is illustrated.


FIGURE 39. Resource request and allocation on the RACH and FACH respectively, followed by transmission of data on the dedicated channel.

This method of first requesting for resources before transmitting data is used in caseswhen the MS has large packets to transmit. The overhead caused by the reservation mech-anism is then negligible. Due to the fact that the MS gets assigned a dedicated channel,data transfer will be more reliable than when it would have been transmitted on theRACH. The reason for this is that the dedicated channel is not a shared channel, thus nocollisions are possible, and the MS uses closed-loop power control on the dedicated chan-nel, whereas this is not the case on the RACH.

The reason of having been assigned a set of TF’s and not only one is that the TF can be

changed during transmission. This can be useful for interference control. This change is

done by means of a TF_Change message, which contains the new TF to be used. The

TF_Change message is transmitted on the DCH.

The third method of transmitting packets, illustrated in figure 40 [45], is when the MS

already has a dedicated channel at its disposal.

FIGURE 40. Packet transmission on the dedicated channel.

The MS can then either transmit first a Cap_Req message on the DCH, in case when the

MS has a large amount of data to transmit, or it can just start transmitting, in case the MS

has just a small amount of data to transmit. The MS can already have a DCH at its disposal

due to the fact that it uses it for another service. Another reason can be that the MS having

just finished transmitting packets on the DCH, will then keep the DCH for a certain time.

If in this time new packets arrive, the MS may immediately start transmission, using the

TF that was used during the last data transmission. Between packets on the DCH, link


maintenance is done by sending pilot bits and power-control commands, ensuring that thepacket transmission is spectrally efficient.

If no packets have been generated for too long of a time, the MS will release the DCH.However, the MS will keep the TF set allocated in the Res_All message. Thus, when it hasnew packets to transmit, it only transmits a short-capacity request (Cap_Req) message onthe RACH, after which it may receive a Cap_All message.

2) Real-Time Services: For real-time services, the allocation procedures are very similar.Once a MS has data to transmit, it first requests resources with a Res_Req message. Thisis done on the RACH, or, in the case where the MS already uses a DCH, it is transmittedon the DCH. As a result, the network now allocates the required resources, again by meansof a set of TF’s. In contrast to the packet data case, where the MS first waits for a Cap_All

message, the MS immediately starts transmitting after it has received a Res_All message.

Another difference from the packet data transmission is that the MS is now allowed to use

any TF allocated in the Res_All message. In this way, the MS can support variable bit rate

services such as speech, but also in this case the network can limit the capacity of the MS.

By means of a resource limit message (Res_Limit), it can limit the previously allocated TF

set. The consequence of this message is that a MS now may only use the TF’s out of the

limited TF set. If later on the capacity in the system is sufficient, determined, and indi-

cated to the MS by the network, the MS is allowed to transmit with all TF’s allocated in

the TF set.

3) Mixed Services: The MAC should also be able to support multiple services. As men-

tioned previously, the physical layer is capable of multiplexing bit streams originating

from different services. The MAC protocol controls this process by controlling the data

stream delivered to the physical layer over the TrCH’s. This control can particularly be

important in the case of when there is a lack of capacity in the system.

If a MS wants to transmit data of different services like, for example, a real-time service

such as speech and a packet data service, then it has been assigned two sets of TF’s. One

set is assigned for the real-time service and one for the packet data service. As mentioned,

in the single-service case, the MS may use any TF assigned for the real-time service,

whereas it may only use one of the TF’s of the TF set for the data service. In the multiple-

service case, the MS may use any TF assigned to it for the speech service. In addition, the

MS gets assigned a specific output power/rate threshold. The aggregate rate of both ser-

vices must be below this threshold. The TF’s used for the data service are chosen out of

the allocated TF set in such a way that the aggregate output power/rate will never exceed

the threshold. Thus, the TF’s used for the data service fluctuate adaptively to the used

TF’s of the speech service.

6.9 Radio Network Aspects

One major benefit of CDMA is the avoidance of frequency planning. Nevertheless, cur-

rent second-generation narrowband CDMA systems have proven to be difficult to plan,

mainly due to power planning. Thus, a great deal of effort has been put into reducing the

network planning for WCDMA.


• For WCDMA, power planning is less demanding, as the common downlink channelsare assigned only about 10% of the base station’s output power, a factor three less than

for current narrow-band CDMA systems.

• Relative comparison of base stations’ signal strengths results in easier planning of soft/

softer handover zones, especially in areas covered by cells with different sizes.

• The cell coverage of CDMA is, in general, very dependent on the cell load. AsWCDMA employs admission control and congestion control, the load can be con-trolled and, thus, also the coverage. In the following sections, the benefits and featuresof admission and congestion control will be described.

To further reduce deployment cost, the WCDMA system is designed so that a reuse of oldsecond-generation sites, e.g., GSM sites, is possible. WCDMA link budgets show that acoverage greater than that of GSM 1800 can be achieved for voice users. Furthermore, thelink budgets also show that 384-kbps packet data services can be provided by WCDMAwith the same coverage as voice service for GSM 1800. Consequently, a WCDMA systemsupporting wide-area coverage up to 384-kbps packet data can be deployed using onlyalready existing GSM 1800 sites.

6.9.1 Admission Control

Admitting a new call will always increase the interference level in the system. This inter-ference increase will reduce the cell coverage, so-called cell breathing. In order to securethe cell coverage when the load increases, the admission control will limit the interfer-ence. The basic strategy is to protect ongoing calls by denying a new user access to thesystem if the system load is already high since dropping is assumed to be more annoyingthan blocking.

In a highly loaded system, the interference increase may cause the system to enter anunstable state and may lead to call dropping. Hence, in addition to securing cell coverage,the admission control is used in order to achieve high capacity and still maintain systemstability.

Admission control is required in both links, since the system is capable of serving differ-ent services. Furthermore, different services demand different capacity as well as differentquality. Hence, service-dependent admission control thresholds will be employed. Theseservices-dependent thresholds should preferably depend on load estimates, for instance,the received power level at the base station as an uplink load estimate and the total trans-mitted power from a base station as a downlink load estimate. Since the received powerlevel as well as the transmitted power level may change rapidly, event-driven measuringand signaling are preferred. The measurement values are obtained at the base station,where the admission decision ought to be made, unless global information is required.Arrivals of high-bit-rate users, particularly the ones that require a large amount ofresources in the downlink, may demand global information in order to make an efficientadmission decision.


6.9.2 Congestion Control

Even though an efficient admission control algorithm and an efficient scheduling proce-dure are employed, an overloaded situation may still occur. When reaching overload, theoutput powers are rapidly increased by the fast closed-loop power control until one or sev-eral transmitters are using their maximum output power. The connections unable toachieve their required quality are considered useless and are only adding interference tothe system. This is, of course, an unacceptable behavior. Hence, a procedure to remove thecongestion is needed. The congestion problem is particularly severe in the uplink, wherethe high-interference levels may propagate in the system. The impact of the high-uplink-interference level, due to overload, may be limited by integrating the uplink power controlwith the uplink congestion control procedure.

This is achieved by slightly degrading the quality of the users in the overloaded cell duringthe time it takes to resolve the congestion. The congestion control consists of severalsteps:

• lowering the bit rate of one or several services that are insensitive to increased delays—

this is the most preferred method;

• performing interfrequency handovers;

• removing one or several connections.

The congestion control is activated once the congestion threshold is exceeded. Thus, boththe uplink and downlink thresholds correspond to a certain load. This means that the samemeasurements as in the admission control are used. However, to detect overload, thesemeasurements have to be updated continuously since the considered values vary very rap-idly when overload occurs. In order to make an efficient decision regarding which connec-tions to deal with, i.e., minimizing the number of altered connections, the congestioncontrol algorithm is likely to require global information. This information is obtained byevent-driven signaling, triggered by the occurrence of overload. Once the connections toalter are identified, the required signaling is typically the same as for altering bit rates, per-forming an interfrequency handover or call termination.

6.10 Performance

Both link and system level performance of WCDMA were investigated during the UMTSradio-interface evaluation phase in ETSI. In this section, some results from this evalua-tion, that are taken from [45], are presented. More simulation results and detailed descrip-tions of simulation assumptions are found in [50] and [53]. The document [54] sets therules for the evaluation and describes the environments more in detail. The evaluation isbased on the REVAL procedure so, e.g., the channel models can be found also in [55].

Dynamic system simulations were performed to translate the link-level results into systemcapacity and spectrum efficiency. The capacity was measured at the point where 98% ofthe users were satisfied (see [54] for definition).


The BER performance of uplink 8-kbps speech is plotted in figure 41 [45] for the outdoor-to-indoor and pedestrian A (3 km/h) and vehicular A (120 km/h) environments. Two-branch antenna diversity is assumed.

FIGURE 41. Link-level BER performance for 8-kbps speech.

The speech data is coded using a rate 1/3 convolutional code and interleaved over 20 ms.No TFI is transmitted, i.e., blind-rate detection is assumed. It can be seen that the required

Eb/No to obtain a BER of 10-3 is 3.3 and 5.0 dB for the pedestrian and vehicular environ-ments, respectively. The Eb/No values include all overhead, such as 8-b CRC, 8-b encodertail, and the entire DPCCH. The corresponding downlink Eb/No values, where no antennadiversity is used, are 6.7 and 7.6 dB, respectively.

The system simulations for speech assume 50% voice activity. A Manhattan-like model isused for the pedestrian channel, while the vehicular channel is assumed in a classic three-sector macrocell environment. In the Manhattan environment, the spectrum efficiency ofspeech is 189 kbps/MHz/cell in the uplink, and 163 kbps/MHz/cell in the downlink. Thecorresponding figures for the macrocell environment are 98 and 78 kbps/MHz/cell,respectively. In the macrocell, one cell is the same as one sector.

Unconstrained delay data (UDD) services, i.e., packet services, were also evaluated inETSI. The UDD packet services have characteristics modeling WWW browsing sessionsand are defined in [“Selection procedures for the choice of radio transmission technolo-

gies of the UMTS (UMTS 30.03),” ETSI Tech. Rep. 101 112, version 3.1.0, Nov. 1997].

One packet service defined is the UDD 384 service. In this service, when packets arrive

for transmission over the radio interface, the average bit rate of those packets is 384 kbps.

However, packets do not arrive continuously. The time between packets can be used for

transmission over the radio interface. This means that the average link level bit rate can be

lower than 384 kbps and still meet the requirements. In WCDMA simulations performed,

the link level bit rate used for the UDD 384 service is 240 kbps (the minimum rate allowed

according to [54] is 38.4 kbps).


The UDD 384 service was simulated for the outdoor to indoor and pedestrian A (3 km/h)channel in the Manhattan environment. Rate 1/2 convolutional coding is used to obtain the240-kbps link level bit rate, with interleaving over 10 ms.

In the link-level simulations, the block error rate (BLER) performance for 300-b blockswas studied. A BLER of around 10% is a good working point for the ARQ scheme [20].The 10% BLER value is reached at an value of 0.2 dB in the uplink with antenna diversity(see figure 42 taken from [45]). In the downlink without antenna diversity, the corre-sponding value is 3.2 dB. As for speech, the values include all overhead, including a TFIfield.

FIGURE 42. Link-level BER and BLER performance for 240-kbps data.

In the system simulations, an ARQ scheme with retransmissions of 300 bit blocks wasused to find the UDD 384 performance. Simulations of the UDD 384 packet service, usingthe 240-kbps link, show a spectrum efficiency in the uplink and downlink of 470 and 565kbps/MHz/cell, respectively.

Using a 384-kbps link level bit rate will lead to similar spectrum efficiency numbers.WCDMA spectrum efficiency numbers are summarized below in table 7.

TABLE 7. WCDMA Spectral Efficiency for Different Services and Environments

Spectrum efficiency [kbps/MHz/cell]

Service and env. Uplink Downlink

Speech

Manhattan 189 163

Macrocell 98 78

UDD 384

Manhattan 470 565


REFERENCES

[30] Ishida, Y., "Recent Study on Candidate Radio Transmission Technology for IMT-2000," First Annual CDMA European Congress, London, UK, Oct. 1997.

[31] Ojanpera, T., M. Gudmundson, P. Jung, J. Sköld, R. Pirhonen, G. Kramer, and A.

Toskala, "FRAMES ( Hybrid Multiple Access Technology," Proc. of ISSSTA'96, Vol. 1,

Mainz, Germany, Sept. 1996, pp. 320 - 324.

[32] Ojanpera, T., P.O. Anderson, J. Castro, L. Girard, A. Klein, and R. Prasad, "A Com-

parative Study of Hybrid Multiple Access Schemes for UMTS," Proc. of ACTS Mobile

Summit Conf., Vol. 1, Granada, Spain, Nov. 1996, pp. 124 - 130.

[33] Ojanpera, T., J. Sköld, J. Castro, L. Girard, and A. Klein, "Comparison of Multiple

Access Schemes for UMTS," Proc. of VTC'97, Vol. 2, Phoenix, Arizona, USA, May

1997, pp. 490 - 494.

[34] Ojanpera, T., A.Klein, and P.-O. Anderson, "FRAMES Multiple Access for UMTS,"

IEE Colloquium on CDMA Techniques and Applications for Third Generation Mobile

Systems, London, May 1997.

[35] Ovesj, F., E. Dahlman, T. Ojanpera, A. Toskala, and A. Klein, "FRAMES Multiple

Access Mode 2 ( Wideband CDMA," Proc. of PIMRC97, Helsinki, Finland, Sept. 1997,

pp. 42 46.

[36] Nikula, E., A. Toskala, E. Dahlman, L. Girard, and A. Klein, "FRAMES Multiple

Access for UMTS and IMT-2000," EEE Pers. Commun., April 1998.

[37] CSEM/Pro Telecom, Ericsson, France TElEcom (CNET, Nokia, Siemens, "FMA -

FRAMES Multiple Access A Harmonized Concept for UMTS/IMT-2000," ITU Work-

shop on Radio Transmission Technologies for IMT-2000, Toronto, Canada, Sept. 10 - 11,

1997.

[38] ARIB FPLMTS Study Committee, "Report on FPLMTS Radio Transmission Tech-

nology SPECIAL GROUP, (Round 2 Activity Report)," Draft v.E1.1, Jan. 1997.

[39] Adachi, F., K. Ohno, M. Sawahashi, and A. Higashi, "Multimedia mobile radio

access based on coherent DS-CDMA," Proc. of 2nd International workshop on Mobile

Multimedia Commun., A2.3, Bristol University, UK Apr. 1995.

[40] Ohno, K., M. Sawahashi, and F. Adachi, "Wideband coherent DS-CDMA," Proc. of

VTC' 95, Chicago, Illinois, USA, July 1995, pp. 779 783.

[41] Dohi, T., Y. Okumura, A. Higashi, K. Ohno, and F.Adachi, "Experiments on Coher-

ent Multicode DS-CDMA," Proc. of VTC'96, Atlanta, Georgia, USA, April 1996, pp. 889

893.


[42] Adachi, F., M. Sawahashi, and K. Ohno, "Coherent DS-CDMA: Promising MultipleAccess for Wireless Multimedia Mobile Communications," Proc. of ISSSTA’96, Mainz,Germany, Sept. 1996, pp. 351 358.

[43] Onoe, S., K. Ohno, K. Yamagata and T. Nakamura, "Wideband-CDMA Radio Con-

trol Techniques for Third Generation Mobile Communication Systems," Proc. of VTC97,

Vol. 2, Phoenix, Arizona, USA, May 1997, pp. 835 839.

[44] 3rd Generation Partnership Project; Technical Specification Group RAN; Radio

Interface Protocol Architecture (3G TS 25.301 version 3.2.0).

[45] Dahlman, E., Beming, P., Knutsson, J., Ovesjo, F., Persson, M., and Roobol, C.,

“WCDMA - The Radio Interface for Future Mobile Multimedia Communications”, IEEE

Transactions on Vehicular Technology, Vol. 47, No 4, November 1998, pp. 1105 - 1118.

[46] 3rd Generation Partnership Project; Technical Specification Group Radio Access

Network; Physical channels and mapping of transport channels onto physical channels

(FDD) (3G TS 25.211 version 3.0.0).

[47] Toskala, A., E. Dahlman, M. Gustafsson, M. Latva-Aho, and M. J. Rinne, "Frames

Multiple Access Mode 2 Physical Transport Control Functions," Proc. of ACTS Summit,

Aalborg, Denmark, Oct. 1997, pp. 697 702.

[48] Higuchi, K., M. Sawahashi and F. Adachi, "Fast Cell Search Algorithm in DS-

CDMA Mobile Radio Using Long Spreading Codes," Proc. of VTC97, Vol.3, Phoenix,

Arizona, USA, May 1997, pp. 1430 - 1434.

[49] R. Esmailzadeh and M. Gustafsson, “A new slotted ALOHA based random access

method for CDMA systems”, in Proc. IEEE Int. Conf. on Universal Personal Commun.,

ICUPC’97, San Diego, CA, Oct. 1997, pp. 43 - 47.

[50] “UMTS terrestrial radio access: Concept evaluation (UMTS 30.06)”, ETSI Tech.

Rep. 101 146, version 3.0.0, Dec. 1997.

[51] C. Roobol, P. Beming, J. Lundsjo, and M. Johansson, "A proposal for an RLC/MAC

protocol for wideband CDMA capable of handling real time and non real time services,"

in Proc. 48th IEEE Veh. Technol. Conf., VTC’98, Ottawa, Canada, May 1998, pp. 107 -

111.

[52] J. Cai and D.J. Goodman, “General packet radio service in GSM,” IEEE Commun.

Mag., vol. 35, pp. 122 - 131, Oct. 1997.

[53] G. Brismark, H. Olofsson, J. Knutsson, M. Gustafsson, and E. Dahlman, “UMTS/

IMT2000 based on wideband CDMA,” in Proc. Multi-Dimensional Mobile Communica-

tions, MDMC’98, Sept. 1998, pp. 18 - 23.

[54] “Selection procedures for the choice of radio transmission technologies of the UMTS

(UMTS 30.03),” ETSI Tech. Rep. 101 112, version 3.1.0, Nov. 1997.


[55] “Guidelines for evalution of radio transmision technologies for IMT2000/FPLMTS,”

ITU-R M.1225 (REVAL).

wcdma and cdma2000 - the radio interfaces for future

Documents