wcdma and cdma2000 - the radio interfaces for future...

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

WCDMA and cdma2000 - The Radio Interfaces for Future Mobile Multimedia

Communications - Part I

Emre A. Yavuz

1.0 Introduction

Mobile telephony is coming to play the lead role in the communications business. Differ-ent markets such as computer, audio and video are converging to the point that data andmultimedia users are beginning to demand that these services are available on the move.For more than a decade, research has been going on to find enabling techniques to intro-duce multimedia capabilities into mobile communications. Research efforts have beenaligned with efforts in the International Telecommunication Union (ITU) and other bodiesto find standards and recommendations, which ensure that the mobile communications ofthe future have access to multimedia capabilities and service quality similar to the fixednetwork.

The European Commission has sponsored research programs such as Research and Devel-opment of Advanced Communications Technologies in Europe, RACE-1 and RACE-2,and Advanced Communications Technology and Services (ACTS) in order to stimulateresearch on future mobile communication. In particular the projects CODIT (evaluation ofcode division multiple access, CDMA) and ATDMA (evaluation of time division multipleaccess, TDMA) within RACE-2 and FRAMES within ACTS have been very important forUniversal Mobile Telecommunications System (UMTS) / International Mobile Telecom-munications in the year 2000 (IMT-2000).

These efforts have been denoted "third generation (3G) mobile communications" in pur-suit of global recommendations and standards for multimedia-capable mobile communi-cations.

First generation mobile communications provided analog voice communications and othertelephony services to mobile users with the standards AMPS, TACS and NMT, whereassecond generation mobile communications provided digital voice communications withdata services as well. The second generation standards, which have mainly circuit-switched low to medium rate data communications (e.g., 9.6 kb/s) are Global System forMobile Communications (GSM), Digital AMPS (DAMPS) / IS-136, Personal Digital Cel-lular (PDC), and cdmaOne/IS-95.

With the introduction of the third generation (UMTS/IMT-2000), second generation capa-bilities (voice and low/medium rate data) are extended adding multimedia capabilities to


second-generation platforms such as support for high bit rates and introduction of packetdata/IP access.

In line with efforts of ITU to provide global recommendations for IMT-2000, a spectrumidentification has been made, identifying parts of the 2 GHz band for IMT-2000 usage.European Telecommunications Standards Institute (ETSI) / Special Mobile Group (SMG),which is one of the standards bodies in the process of making standards for IMT-2000, hasbeen responsible for UMTS standardization since the early 1990s. A historic milestonewas reached in January 1998, when the basic technology for the UMTS terrestrial radioaccess (UTRA) system, that contains the following features, was selected [1].

- For the paired bands 1920-1980 and 2110-2170 MHz wideband CDMA (W-CDMA)shall be used in frequency-division duplex (FDD) operation.

- For the unpaired bands of total 35 MHz time-division code-division multiple access(TD-CDMA) shall be used in time-division duplex (TDD) operation.

- Parameters shall be chosen to facilitate easy implementation of FDD/TDD dual modeterminals.

On the other hand in the United States in March 1998, the TIA (TelecommunicationsIndustry Association) TR45.5 commitee, responsible for IS-95 standardization, adopted aframework for wideband CDMA backward compatible to IS-95, called cdma2000. Origi-nally under TIA TR45.5, the responsibility of the cdma2000 work has been moved to anewly formed organization called “The Third Generation Partnership Project 2,” or

3GPP2, thereby benefiting from the expertise of specialists from Chinese (CWTS), Japa-

nese (ARIB and TTC), Korean (TTA), and North American (TIA) standards bodies. The

ITU approved the cdma2000 radio access system as the CDMA Multi-Carrier (MC) mem-

ber of the IMT-2000 familiy of standards. cdma2000 release 0 was published in the sum-

mer of 1999, release A in late 2000, and release B was supposed to be published by the

end of 2000.

The first part of the report is organized as follows. CDMA technology and its concepts are

presented after the introduction section. A number of modulation techniques that generate

spread-spectrum signals are also discussed and compared with each other regarding the

concepts mentioned in the previous section. It will be followed by another section on basic

elements of DS-CDMA, which is the most popular spread spectrum modulation tecnique.

Then the features of the air interface of IS-95, which is a second generation system already

based on CDMA technology, are going to be described with a focus on the new downlink

and uplink channel structure according to the new IS-95B standard. The following section

will present the origins of WCDMA technology that has been chosen as the basic radio

access technology for UMTS/IMT-2000 in both Europe and Japan and a discussion on

characteristics of WCDMA waveforms, which will lead to a comparison with NCDMA

networks. Signal designs and receiver structures that are being considered for WCDMA

systems will be given and followed by multicarrier and multicode CDMA systems which

differ from those in use in the NCDMA cellular system.


In the second part of the report, the main technical approaches to WCDMA air interfacetechnology are going to be discussed in detail as it will be mostly focused on physicallayer structure, associated procedures, MAC, LAC and radio resource management. Thispart will be followed by a third part, in which same features will be discussed forcdma2000 air interface technology.

2.0 CDMA Technology and Its Concepts

The origins of spread spectrum are in military field and navigation systems. Techniquesdeveloped to counteract intentional jamming have also proved suitable for communicationthrough dispersive channels in cellular applications. In this section the milestones whichare summarized in [2] for CDMA development starting from the 1950s after the inventionof the Shannon theorem are highlighted [3].

In 1949, John Pierce wrote a technical memorandum where he described a multiplexingsystem in which a common medium carries coded signals that need not be synchronized.This system can be classified as a time hopping spread spectrum multiple access system[4]. Claude Shannon and Robert Pierce introduced the basic ideas of CDMA in 1949 bydescribing the interference averaging effect and the graceful degradation of CDMA [5]. In1950, De Rosa-Rogoff proposed a direct sequence spread spectrum system and introducedthe processing gain equation and noise multiplexing idea [4]. In 1956, Price and Greenfiled for the antimultipath "RAKE" patent [4]. Signals arriving over different propagationpaths can be resolved by a wideband spread spectrum signal and combined by the RAKEreceiver. The near-far problem (i.e., a high interference overwhelming a weaker spreadspectrum signal) was first mentioned in 1961 by Magnuski [4].

For cellular application spread spectrum was suggested by Cooper and Nettleton in 1978[6]. During the 1980s Qualcomm investigated DS-CDMA techniques, which finally led tothe commercialization of cellular spread spectrum communications in the form of the nar-rowband CDMA IS-95 standard in July 1993. Commercial operation of IS-95 systemsstarted in 1996. Multiuser detection (MUD) has been subject to extensive research since1986 when Verdu formulated an optimum multiuser detection for the additive white Gaus-sian noise (AWGN) channel, maximum likelihood sequence estimator (MLSE) [7].

In CDMA each user is assigned a unique code sequence it uses to encode its information-bearing signal. The receiver, knowing the code sequences of the user, decodes a receivedsignal after reception and recovers the original data. This is possible since the crosscorre-lations between the code of the desired user and the codes of the other users are small.Since the bandwidth of the code signal is chosen to be much larger than the bandwidth ofthe information-bearing signal, the encoding process enlarges (spreads) the spectrum ofthe signal and is therefore also known as spread-spectrum modulation. The resulting sig-nal is also called a spread-spectrum signal, and CDMA is often denoted as spread-spec-trum multiple access (SSMA) [810, 1112].

The spectral spreading of the transmitted signal gives to CDMA its multiple access capa-

bility. It is therefore important to know the techniques necessary to generate spread-spec-


trum signals and the properties of these signals. A spread-spectrum modulation techniquemust be fulfill two criteria: The transmission bandwidth must be much larger than theinformation bandwidth. The resulting radio-frequency bandwidth is determined by a func-tion other than the information being sent (so the bandwidth is statistically independent ofthe information signal). This excludes modulation techniques like frequency modulation(FM) and phase modulation (PM).

The ratio of transmitted bandwidth to information bandwidth is called the processing gain,Gp, of the spread-spectrum, Gp = Bt / Bi, where Bt is the transmission bandwidth and Bi isthe bandwidth of the information-bearing signal.

The receiver correlates the received signal with a synchronously generated replica of thespreading code to recover the original information-bearing signal. This implies that thereceiver must know the code used to modulate the data.

Because of the coding and the resulting enlarged bandwidth, SS signals have a number ofproperties that differ from the properties of narrowband signals. From the communicationsystems point of view, the most interesting ones, that are mentioned in [2], are givenbelow.

Multiple Access Capability -- If multiple users transmit a spread-spectrum signal at thesame time, the receiver will still be able to distinguish between the users provided eachuser has a unique code that has a sufficiently low cross-correlation with the other codes.Correlating the received signal with a code signal from a certain user will then onlydespread the signal of this user, while the other spread-spectrum signals will remainspread over a large bandwidth. Thus, within the information bandwidth the power of thedesired user will be larger than the interfering power provided there are not too manyinterferers, and the desired signal can be extracted.

Protection Against Multipath Interference -- In a radio channel there is not just one pathbetween a transmitter and receiver. Due to reflections (and refractions) a signal will bereceived from a number of different paths. The signals of the different paths are all copiesof the same transmitted signal but with different amplitudes, phases, delays, and arrivalangles. Adding these signals at the receiver will be constructive at some of the frequenciesand destructive at others. In the time domain, this results in a dispersed signal. Spread-spectrum modulation can combat this multipath interference; however, the way in whichthis is achieved depends very much on the type of modulation used.

Privacy -- The transmitted signal can only be despread and the data recovered if the codeis known to the receiver.

Interference Rejection -- Cross-correlating the code signal with a narrowband signal willspread the power of the narrowband signal thereby reducing the interfering power in theinformation bandwidth.

Anti-Jamming Capability, Especially Narrowband Jamming -- This is more or less thesame as interference rejection except the interference is now willfully inflicted on the sys-


tem. It is this property, together with the next one, that makes spread-spectrum modulationattractive for military applications.

Low Probability of Interception (LPI) -- Because of its low power density, the spread-spectrum signal is difficult to detect and intercept by a hostile listener.

There are a number of modulation techniques that generate spread-spectrum signals. Mostimportant ones are going to be briefly discussed in the following subsections [2].

2.1 Direct Sequence

In DS-CDMA the modulated information-bearing signal (the data signal) is directly mod-ulated by a digital, discrete-time, discrete-valued code signal. The data signal can be eitheranalog or digital; in most cases it is digital. In the case of a digital signal the data modula-tion is often omitted and the data signal is directly multiplied by the code signal and theresulting signal modulates the wideband carrier. It is from this direct multiplication thatthe direct sequence CDMA gets its name.

The binary data signal modulates a RF carrier. The modulated carrier is then modulated bythe code signal. This code signal consists of a number of code bits called "chips" that canbe either +1 or 1. To obtain the desired spreading of the signal, the chip rate of the code

signal must be much higher than the chip rate of the information signal. For the code mod-

ulation various modulation techniques can be used, but usually some form of phase shift

keying (PSK) like binary phase shift keying (BPSK), differential binary phase shift keying

(D-BPSK), quadrature phase shift keying (QPSK), or minimum shift keying (MSK) is

employed.

The rate of the code signal is called the chip rate; one chip denotes one symbol when refer-

ring to spreading code signals.

After transmission of the signal, the receiver uses coherent demodulation to despread the

SS signal, using a locally generated code sequence. To be able to perform the despreading

operation, the receiver must not only know the code sequence used to spread the signal,

but the codes of the received signal and the locally generated code must also be synchro-

nized. This synchronization must be accomplished at the beginning of the reception and

maintained until the whole signal has been received. The code synchronization/tracking

block performs this operation. After despreading a data modulated signal results, and after

demodulation the original data can be recovered.

The most four important properties; the multiple access capability, the multipath interfer-

ence rejection, the narrowband interference rejection, and secure/private communication,

the LPI, that are mentioned above are going to be discussed from the view point of

CDMA for each case of modulation technique.

- Multiple access: If multiple users use the channel at the same time, there will be multiple

DS signals overlapping in time and frequency. At the receiver coherent demodulation is

used to remove the code modulation. This operation concentrates the power of the desired


user in the information bandwidth. If the crosscorrelations between the code of the desireduser and the codes of the interfering users are small, coherent detection will only put asmall part of the power of the interfering signals into the information bandwidth.

- Multipath interference: If the code sequence has an ideal autocorrelation function, thenthe correlation function is zero outside the interval [Tc,Tc], where Tc is the chip duration.

This means that if the desired signal and a version that is delayed for more than 2Tc are

received, coherent demodulation will treat the delayed version as an interfering signal,

putting only a small part of the power in the information bandwidth.

- Narrowband interference: The coherent detection at the receiver involves a multiplica-

tion of the received signal by a locally generated code sequence. However, at the transmit-

ter, multiplying a narrowband signal with a wideband code sequence spreads the spectrum

of the narrowband signal so that its power in the information bandwidth decreases by a

factor equal to the processing gain.

- LPI: Since the whole signal spectrum is used by the direct sequence signal all the time, it

will have a very low transmitted power per hertz which makes it very difficult to detect a

DS signal.

Apart from the above-mentioned properties, DS-CDMA has a number of other specific

properties that can be divided into advantageous (+) and disadvantageous (-) behavior:

(+) The generation of the coded signal is easy. It can be performed by a simple multiplica-

tion.

(+) Since only one carrier frequency has to be generated, the frequency synthesizer (car-

rier generator) is simple.

(+) Coherent demodulation of the DS signal is possible.

(+) No synchronization among the users is necessary.

() It is difficult to acquire and maintain the synchronization of the locally generated code

signal and the received signal. Synchronization has to be kept within a fraction of the chip

time.

() For correct reception the synchronization error of locally generated code sequence and

the received code sequence must be very small, a fraction of the chip time. This combined

with the nonavailability of large contiguous frequency bands practically limits the band-

width to 1020 MHz.

() The power received from users close to the base station is much higher than that

received from users further away. Since a user continuously transmits over the whole

bandwidth, a user close to the base will constantly create a lot of interference for users far

from the base station, making their reception impossible. This near-far effect can be

solved by applying a power control algorithm so that all users are received by the base sta-

tion with the same average power. However this control proves to be quite difficult.


2.2 Frequency Hopping

In frequency hopping CDMA, the carrier frequency of the modulated information signal isnot constant but changes periodically. During time intervals T, the carrier frequencyremains the same, but after each time interval the carrier hops to another (or possibly thesame) frequency. The hopping pattern is decided by the code signal. The set of availablefrequencies the carrier can attain is called the hop-set.

The frequency occupation of an FH-SS system differs considerably from a DS-SS system.A DS system occupies the whole frequency band when it transmits, whereas an FH systemuses only a small part of the bandwidth when it transmits, but the location of this part dif-fers in time. On average, both systems will transmit the same power in the frequency band.

The block diagram for an FH-CDMA system is given in figure 1. The data signal is base-band modulated. Using a fast frequency synthesizer that is controlled by the code signal,the carrier frequency is converted up to the transmission frequency.

FIGURE 1. Block diagram of an FH-CDMA transmitter and receiver

The inverse process takes place at the receiver. Using a locally generated code sequence,the received signal is converted down to the baseband. The data is recovered after (base-band) demodulation. The synchronization/tracking circuit ensures that the hopping of thelocally generated carrier synchronizes to the hopping pattern of the received carrier so thatcorrect despreading of the signal is possible.

Within frequency hopping CDMA a distinction is made that is based on the hopping rateof the carrier. If the hopping rate is (much) greater than the symbol rate, one speaks of afast frequency hopping (F-FH). In this case the carrier frequency changes a number oftimes during the transmission of one symbol, so that one bit is transmitted in different fre-quencies. If the hopping rate is (much) smaller than the symbol rate, one speaks of slowfrequency hopping (S-FH). In this case multiple symbols are transmitted at the same fre-quency.

The occupied bandwidth of the signal on one of the hopping frequencies depends not onlyon the bandwidth of the information signal but also on the shape of the hopping signal andthe hopping frequency. If the hopping frequency is much smaller than the informationbandwidth (which is the case in slow frequency hopping), then the information bandwidth


is the main factor that decides the occupied bandwidth. If, however, the hopping frequencyis much greater than the information bandwidth, the pulse shape of the hopping signal willdecide the occupied bandwidth at one hopping frequency. If this pulse shape is very abrupt(resulting in very abrupt frequency changes), the frequency band will be very broad, limit-ing the number of hop frequencies. If it’s made sure that the frequency changes are

smooth, the frequency band at each hopping frequency will be about 1/Th times the fre-

quency bandwidth, where Th is equal to the hopping frequency. Frequency changes can be

made smooth by decreasing the transmitted power before a frequency hop and increasing

it again when the hopping frequency has changed.

Multiple Access -- It is easy to visualize how the F-FH and S-FH CDMA obtain their mul-

tiple access capability. In the F-FH one symbol is transmitted in different frequency bands.

If the desired user is the only one to transmit in most of the frequency bands, the received

power of the desired signal will be much higher than the interfering power and the signal

will be received correctly.

In the S-FH multiple symbols are transmitted at one frequency. If the probability of other

users transmitting in the same frequency band is low enough, the desired user will be

received correctly most of the time. For those times that interfering users transmit in the

same frequency band, error-correcting codes are used to recover the data transmitted dur-

ing that period.

Multipath Interference -- In the F-FH CDMA the carrier frequency changes a number of

times during the transmission of one symbol. Thus, a particular signal frequency will be

modulated and transmitted on a number of carrier frequencies. The multipath effect is dif-

ferent at the different carrier frequencies. As a result, signal frequencies that are amplified

at one carrier frequency will be attenuated at another carrier frequency and vice versa. At

the receiver the responses at the different hopping frequencies are averaged, thus reducing

the multipath interference. Since usually noncoherent combining is used, this is not as

effective as the multipath interference rejection in a DS-CDMA system, but it still gives

quite an improvement.

Narrowband Interference -- Suppose a narrowband signal is interfering on one of the hop-

ping frequencies. If there are Gp hopping frequencies (where Gp is the processing gain),

the desired user will (on the average) use the hopping frequency where the interferer is

located 1/Gp percent of the time. The interference is therefore reduced by a factor Gp.

LPI -- The difficulty in intercepting an FH signal lies not in its low transmission power.

During a transmission, it uses as much power per hertz as a continuous transmission. But

the frequency at which the signal is going to be transmitted is unknown, and the duration

of the transmission at a particular frequency is quite small. Therefore, although the signal

is more readily intercepted than a DS signal, it is still a difficult task to perform.

Apart from the above-mentioned properties, the FH-CDMA has a number of other specific

properties that we can divide into advantageous (+) and disadvantageous (-) behavior:


(+) Synchronization is much easier with FH-CDMA than with DS-CDMA. With FH-CDMA synchronization has to be within a fraction of the hop time. Since spectral spread-ing is not obtained by using a very high hopping frequency but by using a large hop-set,the hop time will be much longer than the chip time of a DS-CDMA system. Thus, an FH-CDMA system allows a larger synchronization error.

(+) The different frequency bands that an FH signal can occupy do not have to be contigu-ous because we can make the frequency synthesizer easily skip over certain parts of thespectrum. Combined with the easier synchronization, this allows much higher spread-spectrum bandwidths.

(+) The probability of multiple users transmitting in the same frequency band at the sametime is small. A user transmitting far from the base station will be received by it even ifusers close to the base station are transmitting, since those users will probably be transmit-ting at different frequencies. Thus, the near-far performance is much better than that ofDS.

(+) Because of the larger possible bandwidth a FH system can employ, it offers a higherpossible reduction of narrowband interference than a DS system.

() A highly sophisticated frequency synthesizer is necessary.

() An abrupt change of the signal when changing frequency bands will lead to an increase

in the frequency band occupied. To avoid this, the signal has to be turned off and on when

changing frequency.

() Coherent demodulation is difficult because of the problems in maintaining phase rela-

tionships during hopping.

2.3 Time Hopping

In time hopping CDMA, the data signal is transmitted in rapid bursts at time intervals

determined by the code assigned to the user. The time axis is divided into frames, and each

frame is divided into M time slots. During each frame the user will transmit in one of the

M time slots. Which of the M time slots is transmitted depends on the code signal assigned

to the user. Since a user transmits all of its data in one, instead of M time slots, the fre-

quency it needs for its transmission has increased by a factor M. A block diagram of a TH-

CDMA system is given in figure 2.


FIGURE 2. Block diagram of a TH-CDMA transmitter and receiver

Figure 3 shows the time-frequency plot of the TH-CDMA systems.

FIGURE 3. Time frequency plot of the TH-CDMA

Comparing figure 3 with figure 4, it is seen that the TH-CDMA uses the whole widebandspectrum for short periods instead of parts of the spectrum all of the time.

FIGURE 4. Time/ frequency occupancy of FH and DS signals

Following the same procedure as for the previous CDMA schemes, the properties of TH-CDMA with respect to multiple access capability, multipath interference rejection, nar-rowband interference rejection, and probability of interception will be discussed.

Multiple access -- The multiple access capability of TH-SS signals is acquired in the samemanner as that of the FH-SS signals; namely, by making the probability of users’ transmis-


sions in the same frequency band at the same time small. In the case of time hopping alltransmissions are in the same frequency band, so the probability of more than one trans-mission at the same time must be small. This is again achieved by assigning differentcodes to different users. If multiple transmissions do occur, error-correcting codes ensurethat the desired signal can still be recovered. If there is synchronization among the users,and the assigned codes are such that no more than one user transmits at a particular slot,then the TH-CDMA reduces to a TDMA scheme where the slot in which a user transmitsis not fixed but changes from frame to frame.

Multipath interference -- In the time hopping CDMA, a signal is transmitted in reducedtime. The signaling rate, therefore, increases and dispersion of the signal will now lead tooverlap of adjacent bits. Therefore, no advantage is to be gained with respect to multipathinterference rejection.

Narrowband interference -- A TH-CDMA signal is transmitted in reduced time. Thisreduction is equal to 1/Gp, where Gp is the processing gain. At the receiver only an inter-fering signal will be received during the reception of the desired signal. Thus, only 1/Gppercent of the time, which will reduce the interfering power by a factor Gp, is the interfer-ing signal will be received.

LPI -- With TH-CDMA the frequency at which a user transmits is constant but the times atwhich a user transmits are unknown, and the durations of the transmissions are very short.Particularly when multiple users are transmitting, this makes it difficult for an interceptingreceiver to distinguish the beginning and end of a transmission and to decide which trans-missions belong to which user.

Apart from the above-mentioned properties, the TH-CDMA has a number of other spe-cific properties that can be divided into advantageous (+) and disadvantageous (-) behav-ior:

(+) Implementation is simpler than that of FH-CDMA.

(+) It is a very useful method when the transmitter is average-power limited but not peak-power limited since the data are transmitted is short bursts at high power.

(+) As with the FH-CDMA, the near-far problem is much less of a problem since TH-CDMA is an avoidance system, so most of the time a terminal far from the base stationtransmits alone, and is not hindered by transmissions from stations close by.

() It takes a long time before the code is synchronized, and the time in which the receiver

has to perform the synchronization is short.

() If multiple transmissions occur, a large number of data bits are lost, so a good error-cor-

recting code and data interleaving are necessary.


2.4 Hybrid Systems

The hybrid CDMA systems include all CDMA systems that employ a combination of twoor more of the above-mentioned spread-spectrum modulation techniques or a combinationof CDMA with some other multiple access technique. By combining the basic spread-spectrum modulation techniques, four possible hybrid systems: DS/FH, DS/TH, FH/TH,and DS/FH/TH can be formed; and by combining CDMA with TDMA or multicarriermodulation two more can be formed: CDMA/TDMA and MC-CDMA.

The idea of the hybrid system is to combine the specific advantages of each of the modu-lation techniques. For example, if the combined DS/FH system is considered, the advan-tage of the anti-multipath property of the DS system will be combined with the favorablenear-far operation of the FH system, although a disadvantage will lie in the increased com-plexity of the transmitter and receiver. For illustration purposes, a block diagram of a com-bined DS/FH CDMA transmitter is given in figure 5.

FIGURE 5. Hybrid DS-FH transmitter

The data signal is first spread using a DS code signal. The spread signal is then modulatedon a carrier whose frequency hops according to another code sequence. A code clockensures a fixed relation between the two codes.

3.0 Basic DS-CDMA Elements

In this section, the fundamental elements for understanding direct sequence CDMA and itsapplication into third-generation systems, namely, RAKE receiver, power control, softhandover, interfrequency handover, and multiuser detection are reviewed.

3.1 Rake Receiver

A spread-spectrum signal waveform is well matched to the multipath channel. In a multi-path channel, the original transmitted signal reflects from obstacles such as buildings, andmountains, and the receiver receives several copies of the signal with different delays. Ifthe signals arrive more than one chip apart from each other, the receiver can resolve them.Actually, from each multipath signal’s point of view, other multipath signals can be


regarded as interference and they are suppressed by the processing gain. However, a fur-ther benefit is obtained if the resolved multipath signals are combined using RAKEreceiver. Thus, the signal waveform of CDMA signals facilitates utilization of multipathdiversity. Expressing the same phenomenon in the frequency domain means that the band-width of the transmitted signal is larger than the coherence bandwidth of the channel andthe channel is frequency selective (i.e., only part of the signal is affected by the fading).

RAKE receiver consists of correlators, each receiving a multipath signal. After despread-ing by correlators, the signals are combined using, for example, maximal ratio combining.Since the received multipath signals are fading independently, diversity order and thusperformance are improved. Figure 6 illustrates the principle of RAKE receiver.

FIGURE 6. Principle of RAKE receiver

After spreading and modulation, the signal is transmitted and it passes through a multipathchannel, which can be modeled by a tapped delay line (i.e., the reflected signals aredelayed and attenuated in the channel). In figure 6, we have three multipath componentswith different delays (*1, *2, and *3) and attenuation factors (a1, a2, and a3), each corre-sponding to a different propagation path. The RAKE receiver has a receiver finger foreach multipath component. In each finger, the received signal is correlated by a spreadingcode, which is time-aligned with the delay of the multipath signal. After despreading, thesignals are weighted and combined. In figure 6, maximal ratio combining is used, that is,each signal is weighted by the path gain (attenuation factor). Due to the mobile movementthe scattering environment will change, and thus, the delays and attenuation factors willchange as well. Therefore, it is necessary to measure the tapped delay line profile and toreallocate RAKE fingers whenever there is need. Small-scale changes, less than one chip,are taken care of by a code tracking loop, which tracks the time delay of each multipathsignal.

3.2 Power Control

In the uplink of a DS-CDMA system, the requirement for power control is the most seri-ous negative point. The power control problem arises because of the multiple access inter-ference. All users in a DS-CDMA system transmit the messages by using the samebandwidth at the same time and therefore users interfere with one another. Due to thepropagation mechanism, the signal received by the base station from a user terminal close


to the base station will be stronger than the signal received from another terminal locatedat the cell boundary. Hence, the distant users will be dominated by the close user. This iscalled the near-far effect. To achieve a considerable capacity, all signals, irrespective ofdistance, should arrive at the base station with the same mean power. A solution to thisproblem is power control, which attempts to achieve a constant received mean power foreach user. Therefore, the performance of the transmitter power control (TPC) is one of theseveral dependent factors when deciding on the capacity of a DS-CDMA system.

In contrast to the uplink, in the downlink all signals propagate through the same channeland thus are received by a mobile station with equal power. Therefore, no power control isrequired to eliminate near-far problem. The power control is, however, required to mini-mize the interference to other cells and to compensate against the interference from othercells. The worst-case situation for a mobile station occurs when the mobile station is at thecell edge, equidistant from three base stations. However, the interference from other cellsdoes not vary very abruptly.

In addition being useful against interfering users, power control improves the performanceof DS-CDMA against fading channel by compensating the fading dips. If it followed thechannel fading perfectly, power control would turn a fading channel into AWGN channelby eliminating the fading dips completely.

There exist two types of power control principles: open loop and closed loop. The openloop power control measures the interference conditions from the channel and adjusts thetransmission power accordingly. However, since the fast fading does not correlatebetween uplink and downlink, open loop power control will achieve the right power targetonly on average. Therefore, closed loop power control is required. The closed loop powercontrol measures the signal-to-interference ratio (SIR) and sends commands to the trans-mitter on the other end to adjust the transmission power.

3.3 Soft Handover

In soft handover a mobile station is connected to more than one base station simulta-neously. Soft handover is used in CDMA to reduce the interference into other cells and toimprove performance through macro diversity. Softer handover is a soft handover betweentwo sectors of a cell.

Neighboring cells of a cellular system using either FDMA or TDMA do not use the fre-quencies used by the given cell (i.e., there is spatial separation between cells using thesame frequencies). This is called the frequency reuse concept. Because of the processinggain, such spatial separation is not needed in CDMA, and frequency reuse factor of onecan be used. Usually, a mobile station performs a handover when the signal strength of aneighboring cell exceeds the signal strength of the current cell with a given threshold. Thisis called hard handover. Since in a CDMA system the neighboring cell frequencies are thesame as in the given cell, this type of approach would cause excessive interference into theneighboring cells and thus a capacity degradation. In order to avoid this interference, aninstantaneous handover from the current cell to the new cell would be required when thesignal strength of the new cell exceeds the signal strength of the current cell. This is not,


however, feasible in practice. The handover mechanism should always allow the mobilestation to connect into a cell, which it receives with the highest power (i.e., with the lowestpathloss). Since in soft handover the mobile station is connected to either two or morebase stations, its transmission power can be controlled according to the cell, which themobile station receives with the highest signal strength. A mobile station enters the softhandover state when the signal strength of neighboring cell exceeds a certain threshold butis still below the current base station’s signal strength.

Fortunately, the signal structure of CDMA is well suited for the implementation of softhandover. This is because in the uplink, two or more base stations can receive the samesignal because of the reuse factor of one; and in the downlink the mobile station cancoherently combine the signals from different base stations since it sees them as just addi-tional multipath components. This provides an additional benefit called macro diversity(i.e., the diversity gain provided by the reception of one or more additional signals). Aseparate channel called pilot is usually used for the signal strength measurements for han-dover purposes.

In the downlink, however, soft handover creates more interference to the system since thenew base station now transmits an additional signal for the mobile station. It is possiblethat the mobile station cannot catch all the energy that the base station transmits due to alimited number of RAKE fingers. Thus, the gain of soft handover in the downlink dependson the gain of macro diversity and the loss of performance due to increased interference.Figure 7 illustrates the soft handover principle with two base stations involved.

FIGURE 7. Principle of soft handover with two base station transceivers (BTS)

In the uplink the mobile station signal is received by the two base stations, which, afterdemodulation and combining, pass the signal forward to the combining point, typically tothe base station controller (BSC). In the downlink the same information is transmitted viaboth base stations, and the mobile station receives the information from two base stationsas separate multipath signals and can therefore combine them.

3.4 Interfrequency Handover

The third-generation CDMA networks will have multiple frequency carriers in each cell,and a hot-spot cell could have a larger number of frequencies than neigboring cells. Fur-


thermore, in hierarchical cell structures, micro cells will have a different frequency thanthe macro cell overlaying the micro cells. Therefore, an efficient procedure is needed for ahandover between different frequencies. A blind handover used by second-generationCDMA does not result in an adequate call quality. Instead, the mobile station has to beable to measure the signal strength and quality of an another carrier frequency, while stillmaintaining the connection in the current carrier frequency. Since a CDMA transmissionis continuous, there are no idle slots for the interfrequency measurement/ as in the TDMA-based systems. Therefore, compressed mode and dual receiver have been proposed as asolution to interfrequency handover [13]. In the compressed mode, measurements slots arecreated by transmitting the data of a frame, for example, with a lower spreading ratio dur-ing a shorter period, and the rest of the frame is utilized for the measurements on other car-riers. The dual receiver can measure other frequencies without affecting the reception ofthe current frequency.

3.5 Multiuser Detection

The current CDMA receivers are based on the RAKE receiver principle, which considersother users’ signals as interference. However, in an optimum receiver all signals would bedetected jointly or interference from other signals would be removed by subtracting themfrom the desired signal. This is possible because the correlation properties between signalsare known (i.e., the interference is deterministic not random).

The capacity of a direct sequence CDMA system using RAKE receiver is interferencelimited. In practice this means that when a new user, or interferer, enters the network,other users’ service quality will go below the acceptable level. The more the network canresist interference the more users can be served. Multiple access interference that disturbsa base or mobile station is a sum of both intra- and inter-cell interference.

Multiuser detection (MUD), also called joint detection and interference cancellation (IC),provides a means of reducing the effect of multiple access interference, and henceincreases the system capacity. In the first place MUD is considered to cancel only theintra-cell interference, meaning that in a practical system the capacity will be limited bythe efficiency of the algorithm and the inter-cell interference.

In addition to capacity improvement, MUD alleviates the near/far problem typical to DS-CDMA systems. A mobile station close to a base station may block the whole cell trafficby using too high a transmission power. If this user is detected first and subtracted fromthe input signal, the other users do not see the interference.

Since optimal multiuser detection is very complex and in practice impossible to imple-ment for any reasonable number of users, a number of suboptimum multiuser and interfer-ence cancellation receivers have been developed. The suboptimum receivers can bedivided into two main categories: linear detectors and interference cancellation. Lineardetectors apply a linear transform into the outputs of the matched filters that are trying toremove the multiple access interference (i.e., the interference due to correlations betweenuser codes). Examples of linear detectors are decorrelator and linear minimum meansquare error (LMMSE) detectors. In interference cancellation multiple access interference


is first estimated and then subtracted from the received signal. Parallel interference cancel-lation (PIC) and successive (serial) interference cancellation (SIC) are examples of inter-ference cancellation.

4.0 IS-95 CDMA

In this section, the features of the IS-95 air interface according to the new IS-95B stan-dard, with a focus on the new downlink and uplink channel structure are going to bedescribed [14]. Main air interface parameters, downlink and uplink channel structures,power control principles, and speech coding will be discussed. For a more detailed treat-ment of the IS-95A standard, [11] can be referred whereas for a theoretical analysis of IS-95 air interface solutions, [12] is the one to be referred..

The IS-95 air interface standard, after the first revision in 1995, was termed IS-95A [15];it specifies the air interface for cellular, 800-Mhz frequency band. ANSI J-STD-008 spec-ifies the PCS version (i.e., the air interface for 1900 MHz). It differs from IS-95A prima-rily in the frequency plan and in call processing related to subscriber station identity, suchas paging and call origination. TSB74 specifies the Rate Set 2 (14.4 Kb/s) standard. IS-95B merges the IS-95A, ANSI J-STD-008 [16], and TSB74 standards, and, in addition, itspecifies the high-speed data operation using up to eight parallel codes, resulting in a max-imum bit rate of 115.2 Kb/s. In addition to these air interface specifications, the IS-97 [17]and IS-98 [18] standards specify the minimum performance specifications for the mobileand base station, respectively. Table 1, which is shown below, lists the main parameters ofthe IS-95 air interface.

TABLE 1. IS-95 Air interface parameters

Bandwidth 1.25 MHz

Chip Rate 1.2288 Mc/s

Frequency band uplink 869 - 894 MHz

1930 - 1980 MHz

Frequency band downlink 824 - 849 MHz

1850 - 1910 MHz

Frame length 20 ms

Bit rates Rate set 1: 9.6 kb/s

Rate set 2: 14.4 Kb/s

IS-95B:115.2 Kb/s

Speech code QCELP 8Kb/s

ACELP 13 Kb/s

Soft handover Yes

Power control Uplink:open loop + fast closed loop

Downlink: slow quality loop

Number of RAKE fingers 4

Spreading codes Walsh+Long M-sequence


Carrier spacing of the system is 1.25 MHz. Practical deployment has shown that 3 CDMAcarriers can be fitted into 5 MHz bandwidth due to required guard bands. Network is syn-chronous within few microseconds. This facilitates use of the same long code sequencewith different phase offsets as pilot sequences. However, an external reference signal suchas GPS is needed.

4.1 Downlink Channel Structure

The pilot channel, the paging channel, and the synchronization channel (In the IS-95 stan-dard the synchronization channel is actually termed "Sync Channel.") are common controlchannels and traffic channels are dedicated channels. A common channel is a shared chan-nel, and a dedicated channel is solely allocated for the use of a single user. Data to betransmitted on synchronization, paging, and traffic channels are first grouped into 20 msframes, convolutionally encoded, repeated to adjust the data rate, and interleaved. Thenthe signal is spread with an orthogonal Walsh code at a rate of 1.2288 Mc/s, split into the Iand Q channels, and, prior to baseband filtering, spread with long PN sequences at a rateof 1.2288 Mc/s.

A mobile station uses the pilot channel for coherent demodulation, acquisition, time delaytracking, power control measurements, and as an aid for the handover. In order to obtain areliable phase reference for coherent demodulation, the pilot channel is transmitted withhigher power than the traffic channels. Typically about 20 percent of the radiated poweron the downlink is dedicated to the pilot signal. After obtaining phase and code synchroni-zation, the mobile station acquires synchronization information (data rate of the pagingchannel, time of the base station’s pilot PN sequence with respect to the system time) fromthe synchronization channel. Since the synchronization channel frame has the same lengthas the pilot sequence, acquisition of the synchronization channel takes place easily. Thesynchronization channel operates at a fixed rate of 1.2 Kb/s. The paging channel is used topage a mobile station. The paging channel has a fixed data rate of 9.6 or 4.8 Kb/s.

Each forward traffic channel contains one fundamental code channel and may contain oneto seven supplemental code channels. The traffic channel has two different rate sets. Therate set 1 supports data rates of 9.6, 4.8, 2.4, and 1.2 Kb/s and the rate set 2 supports 14.4,7.2, 3.6, and 1.8 Kb/s. Only the full rate (9.6 or 14.4 Kb/s) may be utilized on the supple-mental code channels. The mobile station always supports the rate set 1 and it may supportthe rate set 2. To achieve equal power levels at the base station receiver, the base stationmeasures the received signal and adjusts each mobile station’s power levels accordingly.The 20 ms frame is divided into 16 power control groups with a duration of 1.25 ms. Onepower control bit is multiplexed in for the fundamental code channel for each power con-trol group.

The transmitted data is encoded by a convolutional code with a constraint length of 9. Thegenerator functions for this code are 753 (octal) and 561 (octal). For the synchronizationchannel, the paging channels, and rate set 1 on the traffic channel, a convolutional codewith a rate of 1/2 is used. For the rate set 2, an effective code rate of 3/4 is achieved bypuncturing two out of every six symbols after the symbol repetition.


Since the data rate on different channels varies, symbol repetition is used to achieve afixed data rate prior to interleaving. For the synchronization channel, each convolutionallyencoded symbol shall be repeated once (i.e., each symbol occurs two consecutive times).For the paging channel, each code symbol at the 4800 b/s rate shall be repeated once. Thecode symbol repetition rate on the forward traffic channels varies with data rate. Codesymbols are not repeated for the 14.4 and 9.6 Kb/s data rates. Each code symbol at the 7.2and 4.8 Kb/s data rates is repeated once, at the 3.6 and 2.4 Kb/s data rates three times, andat the 1.8 and 1.2 Kb/s data rates seven times.

In the downlink, three types of spreading codes are used. Walsh codes of length 64 at afixed chip rate of 1.2288 Mc/s separate the physical channels. The Walsh function consist-ing of all zeros W0 (Walsh code number 0) is used for the pilot channel, W1-W7 are usedfor paging channels (unused paging channel codes can be used for traffic channels). Thesynchronization channel is W32, and traffic channels are W8 to W31 and W33 to W63. Apair of long M-sequences of length 16,767 (2151) is used for quadrature spreading, one

for the I channel and one for the Q channel. Quaternary spreading is used to obtain better

interference averaging. Since the pilot channel Walsh function is all zeros, this pair of

sequences also forms the pilot code. Different cells and sectors are distinguished with the

different phase offsets of this code.

A long pseudo random sequence with a period of 2421 is used for base band data scram-

bling (i.e., to encrypt the signal on the paging and traffic channels). It is decimated from a

1.2288 Mc/s rate down to 19.2 Kb/s. The long pseudo noise sequence is the same used in

the uplink for user separation, and it is generated by a modulo-2 inner product of a 42-bit

mask and the 42-bit state vector of the sequence generator.

4.2 Uplink Channel Structure

Uplink has two physical channels: a traffic channel, which is a dedicated channel, and a

common access channel. A traffic channel consists of a single fundamental code channel

and zero through seven supplemental code channels. Similar to the downlink, traffic chan-

nels always support the rate set 1 data rates and may support the rate set 2 data rates. The

supplemental code channel can only use the full rates (9.6 or 14.4 Kb/s). Data transmitted

on the uplink channels are grouped into 20 ms frames, convolutionally encoded, block

interleaved, and modulated by 64-ary orthogonal modulation. Then, prior to baseband fil-

tering, the signal is spread with a long PN sequence at a rate of 1.2288 Mc/s, split into the

I and Q channels, and spread with in-phase and quadrature spreading sequences.

The access channel is used by a mobile station to initiate a call, to respond to a paging

channel message from the base station, and for a location update. Each access channel is

associated with a downlink paging channel, and consequently there can be up to seven

access channels. The access channel supports fixed data rate operation at 4.8 Kb/s.

The transmitted information is encoded using a convolutional code with constraint length

9 and the same generator polynomials as in the downlink. For the access channel and rate

set 1 on the traffic channels, the convolutional code rate is 1/3. For rate set 2 on the traffic

channels, a code rate of 1/2 is used. Similar to the downlink, code symbols output from the


convolutional encoder are repeated before being interleaved when the data rate is lowerthan 9.6 Kb/s for rate set 1 and 14.4 Kb/s for rate set 2. However, the repeated symbols arenot actually transmitted. They are masked out according to a masking pattern generated bythe data burst randomizer to save transmission power. For the access channel, which has afixed data rate of 4.8 Kb/s, each code symbol is repeated once. In contrast to the trafficchannel, the repeated code symbols are transmitted.

The coded symbols are grouped into 6-symbol groups. These groups are then used toselect one of 64 possible Walsh symbols (i.e., a 64-ary orthogonal modulation is carriedout to obtain good performance for noncoherent modulation). After the orthogonal modu-lation, the transmission rate is 307.2 Kb/s. The reason to use the non-coherent modulationis the difficulty in obtaining good phase reference for coherent demodulation in the uplink.It should be noted how the Walsh codes are used differently in the uplink and downlink. Inthe downlink, they were used for channelization, while in the uplink they are used fororthogonal modulation.

Each code channel in a traffic channel and each access channel are identified by a differentphase of a pseudo-random M-sequence with a length of 242. The in-phase and quadraturespreading is performed by the same pair of M-sequences (length 215) as in the downlink(now augmented by one chip).

4.3 Power Control

IS-95 has three different power control mechanisms. In the uplink, both open loop and fastclosed loop power control are employed. In the downlink, a relatively slow power controlloop controls the transmission power.

4.3.1 Open Loop Power Control

The open loop power control has two main functions: it adjusts the initial access channeltransmission power of the mobile station and compensates large abrupt variations in thepathloss attenuation. The mobile station determines an estimate of the pathloss betweenthe base station and mobile station by measuring the received signal strength at the mobileusing an automatic gain control (AGC) circuitry, which gives a rough estimate of the prop-agation loss for each user. The smaller the received power, the larger the propagation loss,and vice-versa. The transmit power of the mobile station is determined from the equation:

mean output power (dBm) = mean input power (dBm) + offset power + parameters

The offset power for the 800-MHz band mobiles (band class 0) is 73 and for the 1900-

MHz band mobiles is (band class 1) 76 [14]. The parameters are used to adjust the open-

loop power control for different cell sizes and different cell effective radiated powers

(ERP) and receiver sensitivities [19]. These parameters are initially transmitted on the

synchronization channel.

The open loop power control principle is described in figure 8.


FIGURE 8. Open loop power control principle

Since the distance (d1) of mobile station 1 to the base station (BTS) is shorter than the dis-tance of mobile station 2 (d2) to the BTS, the signal received by the mobile station 1 has asmaller propagation loss. Assume that the mean input power of the mobile station 1 is 70

dBm (100 pW) [1dBm means 1dB over 1 mW. For example, 70 dBm is 70 dB (10 million

times) less than 1 mW (i.e., 1E-12 W = 1 picoWatt)] and the mean input power of the

mobile station 2 is 90 dBm (1 pW). For band class 0 mobiles with no correction parame-

ters, the mobile station transmission power to achieve equal received powers at the base

station can be calculated from the equation given above to be -3 dBm (500 µW) and 17

dBm (50 mW), respectively.

4.3.2 Closed Loop Power Control

Since the IS-95 uplink and downlink have a frequency separation of 20 MHz, their fading

processes are not strongly correlated. Even though the average power is approximately the

same, the short term power is different, and therefore, the open loop power control cannot

compensate for the uplink fading. To account for the independence of the Rayleigh fading

in the uplink and downlink, the base station also controls the mobile station transmission

power. Figure 9 illustrates the closed loop power control.

FIGURE 9. Closed loop power control principle


The base station measures the received SIR [The IS-95 standard suggests that the receivedsignal strength should be measured. However, in practice usually the SIR or the receivedbit energy to noise density (Eb/Io) are used, since they have direct impact on the bit errorrate (BER).] over a 1.25-ms period, equivalent to six modulation symbols (9.6 kbps),compares that to the target SIR, and decides whether the mobile station transmissionpower needs to be increased or decreased. The power control bits are transmitted on thedownlink fundamental code channel every 1.25 ms (i.e., with a transmission rate of 800Hz, corresponds to every frame 20ms/16 = 1.25ms) by puncturing the data symbols. Theplacement of a power control bit is randomized within the 1.25-ms power control group.The transmission occurs in the second power control group following the correspondinguplink traffic channel power control group in which the SIR was estimated. For instance,if the signal is received on the reverse traffic channel in power control group number 5,and the corresponding power control bit is transmitted on the forward traffic channel dur-ing power control group number 5 + 2 = 7 [14].

Since the power control commands are transmitted uncoded, their error ratio is fairly high,on the order of 5 percent. However, since the loop is of delta modulation type (i.e., poweris adjusted continuously up or down) this is not critical. The mobile station extracts thepower control bits commands and adjusts its transmission power accordingly. The adjust-ment step is a system parameter and can be 0.25, 0.5, or 1.0 dB. The dynamic range for theclosed loop power control is ±24 dB. The composite dynamic range for open and closed

loop power control is ±32 dB for mobile stations operating in band class 0, and ±40 dB for

mobile stations operating in band class 1 [14]. The typical standard deviation of the power

control error due to the closed loop is on the order of 1.1 to 1.5 dB [12].

The SIR required to produce a certain bit error rate varies according to radio environment

and depends on the amount and type of multipath. Therefore, IS-95 employs an outer loop

that adjusts the target SIR. The base station measures the signal quality (bit error rate), and

based on that determines the target SIR. However, this outer loop will increase the power

control error, resulting in a total standard deviation of 1.5 to 2.5 dB [12].

4.3.3 Downlink Slow Power Control

The base station controls its transmission power to a given mobile station according to the

pathloss and interference situation. The main purpose of the slow downlink power control

is to improve the performance of mobile stations at a cell edge where the signal is weak

and the interfering base station signals are strong. The downlink power control mechanism

is as follows. The base station periodically reduces the transmitted power to the mobile

station. The mobile station measures the frame error ratio (FER). When the FER exceeds a

predefined limit, typically 1 percent, the mobile station requests additional power from the

base station. This adjustment occurs every 15 to 20 ms. The dynamic range of the down-

link power control is only ±6 dB.

Both periodic and threshold reporting may be enabled simultaneously, either one of them

may be enabled, or both forms of reporting may be disabled at any given time.


4.4 Speech Codecs and Discontinuous Transmission (DTX)

IS-95 has three speech codecs, 8-Kb/s QCELP, 8-Kb/s EVRC, and 13 Kb/s. The higherrate codec was developed to provide better voice quality, but due to its higher bit rate itreduces the system capacity. Therefore, the enhanced variable rate codec (EVRC) operat-ing at 8 Kb/s was developed. Speech codecs are four-rate code excited linear predictioncodecs (CELP). The vocoder rates of the 8-Kb/s codecs are 1, 2, 4, and 8 Kb/s correspond-ing to channel rates of 1.2, 2.4, 4.8, and 9.6 Kb/s. The 13-Kb/s codec uses a 14.4 Kb/schannel rate. Since the system capacity is directly proportional to interference, reductionof the transmitted data rate results in better capacity. In IS-95, data rate reduction is imple-mented with discontinuous transmission (DTX), which is realized by gating the transmit-ter in pseudo-random fashion on and off. The drawback of this approach is that it createspulsed interference.

5.0 Air Interface Technologies for Third Generation

In the search for the most appropriate multiple access technology for third-generationwireless systems, a number of new multiple access schemes have been proposed (e.g,wideband CDMA schemes, UWC-136 TDMA-based scheme, and TD-CDMA). This sec-tion presents a detailed description of wideband CDMA, the most appropriate one and itsadvantages over narrowband CDMA.

5.1 The origins of wideband CDMA

Spread spectrum communications systems have been in existence for decades, althoughup until the last decade or so, most of the systems have been military systems, where theneed for signals displaying anti-jam and low-probability of intercept characteristics wasparamount. Thus they were typically designed to be wideband, and those that employedDS to achieve multiple access capability were the original forerunners of what is nowcalled wideband CDMA. In the late 1980s, the use of DS CDMA started to becomeincreasingly of interest to the commercial sector for use in cellular type communicationsand both narrowband CDMA and wideband CDMA systems were designed.

Although there is no single, universally accepted definition of wideband CDMA, defini-tions based upon system parameters such as chip rate, or bandwidth as a fraction of centerfrequency can be quite useful in various scenarios. In this report, a definition [20] moreintimately tied to the distinction between wideband and narrowband CDMA when usedover a wireless channel will be used. Specifically, a wideband CDMA system is defined asone wherein the spread bandwidth of the underlying waveforms in the system typicallyexceed the coherence bandwidth of the channel over which the waveforms are transmitted(meaning irrespective of whether the channel is indoors, outdoor urban, outdoor suburban,etc.)

The initial wideband CDMA cellular systems are envisioned to spread over 5 MHz, andfuture wideband CDMA systems are projected to spread over either 10 or 20 MHz. Thereare several reasons for choosing this bandwidth. First, data rates of 144 and 384 Kb/s, the


main targets of third-generation systems, are achievable within 5 MHz bandwidth with areasonable capacity. Even a 2-Mb/s peak rate can be provided under limited conditions.Second, lack of spectrum calls for reasonably small minimum spectrum allocation, espe-cially if the system has to be deployed within the existing frequency bands occupiedalready by second-generation systems. Third, the 5-MHz bandwidth can resolve (separate)more multipaths than narrower bandwidths, increasing diversity and thus improving per-formance. Larger bandwidths of 10, 15, and 20 MHz have been proposed to supporthigher data rates more effectively.

In this section to follow, basic characteristics of wideband CDMA waveforms that makethem attractive for high data rate transmission over wireless and mobile channels are dis-cussed and the effect of spread bandwidth as it relates to comparative performance ofwideband and narrowband systems will be focused [20].

5.1.1 Bit-Error Rate Considerations

It is known that, if one wants a DS system to operate effectively over multipath channel,the channel should appear frequency selective to the waveform; stated differently thespread bandwidth should exceed the coherence bandwidth of the channel allowing a rakereceiver to be employed so that multiple reflections of the transmitted waveform can beresolved and then coherently combined. However, there is a tradeoff in this type ofreceiver design, in that as the spread bandwidth increases, thus allowing more paths to beresolved, the energy per resolved path decreases, thus making it more difficult for thereceiver to estimate the amplitudes and phases of the paths as required by an optimal rakereceiver.

Having been described this tradeoff; it is shown in [21] that when coherent combining isemployed, along with perfect channel estimates, the wideband system always outperformsthe narrowband system. On the other hand, this conclusion does not necessarily follow ifnoncoherent combining is used. To make things more clear, let’s consider first a coherentreceiver, for a wideband system whereby each user spreads its signal over the entire band-width, the number of resolvable paths available for combining is maximized, and thecoherent combining of all those paths results in the most efficient use of the desired sig-nal’s energy. If instead of using the above approach, one uses a hybrid scheme, wherebythe total bandwidth is first divided into smaller segments and those segments individuallysupport narrowband CDMA networks (i.e., if one uses FDMA/CDMA architecture), theultimate benefits of the rake combining are not achieved.

Alternatively, if noncoherent reception is employed with wideband CDMA, as more andmore paths are combined, the system experiences a noncoherent combining loss. Thus,depending on the signal-to-noise ratio and the amount of interference seen by the receiver,there are scenarios for which a hybrid system yields better performance than one, whichspreads across the entire available bandwidth.

Although the results found in [21] are useful, they do not provide a complete picture whenwe return to a coherent system, since they are based upon perfect estimation of the channelparameters. According to the results of the analysis in [22] and [23], in which imperfect


phase estimations are used, as the signal-to-noise ratio (SNR) becomes lower, the abilityto accurately estimate the phase decreases, and thus a tradeoff in overall system perfor-mance exists between wideband CDMA and hybrid FDMA/CDMA, analogous to thetradeoff if noncoherent reception is employed. However one needs a reasonably low SNR[23] in the tracking loop before a hybrid system outperforms a wideband system.

Finally, for a different perspective on the advantage of a wider spread bandwidth, considerthe experimental curves presented in the tutorial paper by Adachi, Sawahashi, and Suda[24]. Results are presented in figures 10 and 11 for the chip rates changing from 0.96Mchips/s to 7.68 Mchips/s. The measured cumulative distribution of the mobile transmitpower is plotted in figure 10.

FIGURE 10. Comparison of mobile transmit powers with different chip rates

The transmit powers are expressed in dB relative to the median (50 percent) transmitpower with 0.96 Mchips/s spreading and no antenna diversity at the cell site. Diversityreception can reduce the mobile transmit power by about 3 dB at the median value. As thechip rate increases, the probability of large transmit powers reduces, because the multipathtime resolution capabilities improve and the fading seen after rake combining becomesshallower. Fast transmit power control compensates for the fading seen after rake combin-ing. As a consequence, the probability of large transmit powers falls as chip rate increases.A similar power reduction as the chip rate increases can also be seen when antenna diver-sity is applied. The measured BER performance is plotted as a function of target Eb/Io atthe rake combiner output in figure 11 with the spreading chip rate as a parameter.

For comparison, the computer simulation results are also plotted for different numbers Lof resolved paths. Without antenna diversity, as the chip rate increases the BER perfor-mance improves; while the BER performance with 0.96 Mchips/s is close to the computersimulated BER performance with L=1, the performance with 7.68 Mchips/s spreadingbecomes almost the same as simulated with L=2. When antenna diversity is used, the BERperformance improvements, achieved by increasing the chip rate are not significant com-pared to those without diversity, as can be expected from the computer simulation. How-ever, it should be pointed out that increasing the chip rate from 0.96 to 7.68 Mchips/s can


reduce the median transmit power by about 4 dB with antenna diversity. It is stated in thesame paper that the results may depend on the surrounding environment; so further mea-surements in different places are required.

FIGURE 11. Comparison of average BER performance with different chip rates

5.1.2 Acquisition

Note that the results referred to in 5.1.1 are based upon bit-error rate considerations, andmost of them correspond to system performance under the assumption of perfect acquisi-tion and tracking of the spreading sequences. The coarse acquisition problem is addressedin [25] and [26], and it is shown that, as is the case if BER is the criterion, widebandspreading provides better performance than does narrowband spreading. Since a widebandDS waveform has multiple correct paths to which one can lock (i.e. if there are, say, Lresolvable multipath components, then there are that many correct phase positions), thelikelihood of observing a correct path during the search process has been increased fromwhat it would be for narrowband CDMA, which corresponds to L=1. To be specific, con-sider [26], in which an optimal strategy is derived for a parallel coarse acquisitionreceiver.

Shown in figure 12 are several sets off curves, where in each set, there is one curve show-ing the performance of a conventional receiver (see figure 2 in [20]) and one showing per-formance of the optimal receiver, corresponding to a test statistic more complex than theconventional one mentioned above (can be found in [26]).

Note that the sets of curves are parameterized by the number of resolvable paths. Interest-ingly, it is seen that, depending on the number of resolvable paths, the conventionalreceiver performs fairly well relative to the optimal receiver. However if one is interestedin enhancing acquisition performance, the more advantageous way to achieve it is toincrease the spread bandwidth in the CDMA system, as opposed to implementing a morecomplex receiver structure at a given spread bandwidth.


FIGURE 12. Comparison of optimum and conventional parallel and acquisition schemes

For example, consider the curves of figure 12 corresponding to two resolvable paths. If the

target false lock probability is, say 10-3, the performance of the optimal receiver is severaltenths of a decibel better than that of the conventional receiver. However, if the spreadbandwidth is doubled, so that now four paths are resolvable, the performance gain some-where between 4 and 5 dB. One could also gain an extra decibel or so of improvement ifsimultaneously the spread bandwidth is doubled and the optimal receiver is used, but mostof the total improvement comes from the use of the wider spread bandwidth.

5.1.3 Power Control

There are two key effects taking place regarding power control and each is favorable to thewideband system. First, the accuracy with which a power measurement can be madeincrease with increasing spread bandwidth, since the amount of fading the signal experi-ence decreased [27]. Second the effect of any residual measurement inaccuracy is less in awideband system than it is in a narrowband system; that is, for a given power control error,the decrease in capacity experienced by a narrowband CDMA system is greater than thatseen by a wideband CDMA system. Even if the power control is perfect, the additionalpower that must be transmitted by a narrowband CDMA system (to overcome theincreased loss due to multipath fading) spills over into surrounding cells, and thus causesincreased intercell interference.

5.1.4 Fading Statistics

When a DS signal is received over a multipath fading channel, the amplitude distributionof the despread signal becomes more and more specular as the spread bandwidthincreases. This is irrespective of the fading statistics of the received waveform. For exam-ple, assume the received signal on any given resolvable path is experiencing Rayleigh fad-ing at the input to a despreader that is synchronized to that path. When the signal is


despread, only a fraction of multiple paths offer any significant contribution to the output(because the remaining paths experience a large attenuation due to the sharply localizedautocorrelation function of a typical spreading sequence). Indeed, as the spread bandwidthincreases, the autocorrelation function becomes more localized, and a larger number ofpaths are attenuated to the extent that they become insignificant. Thus, the despreader out-put becomes dominated by a decreasing number of paths (as spread bandwidth increases),and no longer is characterized by a Rayleigh probability distribution; rather its statisticshave been found to be more closely described by a Rician density. This, of course, sug-gests, that the effect of the estimation errors discussed in 5.1.1 is overly pessimistic, sincethe analyses that produced those results are based upon Rayleigh statistics, independent ofthe spread bandwidth.

5.2 Wideband CDMA Waveform Design Alternatives for Cellular Systems

In this section, some of the signal designs and receiver structures that are being consideredfor wideband CDMA systems will be described.

5.2.1 Coherent Reverse Link

One design change relates to the signal structure on the mobile-to-base link, commonlyreferred to as the reverse link. Noncoherent detection on the reverse link (specifically, theyuse 64-ary orthogonal signaling with noncoherent detection) is the most preferred one fornarrowband CDMA systems, while the current wideband CDMA waveform designs forthe reverse link employ either BPSK or QPSK with coherent detection. This type ofreverse link design is what was proposed for the earlier designs for DS spread spectrumsystems.

5.2.2 Multicarrier CDMA

Another proposed change in the waveform design is a new technique that makes use of amulticarrier (MC) CDMA waveform which has been suggested recently [28]. In such adesign, multiple narrowband DS CDMA waveforms, each at a distinct carrier frequency,are combined to yield a composite wideband CDMA signal. Among the advantages ofsuch an approach is the ability to achieve the same type of system performance that a con-ventional signal carrier, wideband CDMA signal would provide, such as diversityenhancement over a multipath channel; however this is achieved without the need for acontiguous spectral band over which to spread. Thus, regarding the overlay, MC transmis-sion can be especially attractive, since frequency slots occupied by narrowband wave-forms can be avoided altogether by simply not transmitting at the corresponding carrierfrequencies.

As shown in figure 13, for the forward link (i.e. the link from the base to the mobile), thedriving force for such a waveform design is the ease with which one can convert from anarrowband CDMA signal to a wideband CDMA signal.


FIGURE 13. Power spectral densities of single carrier and multicarrier waveforms

That is, by taking, say, M narrowband CDMA waveforms, each on a different carrier fre-quency and assigning them all to one user, one can now increase the spread bandwidth bya factor of M. There are various ways in which one can create an MC signal, and, indeed,there is no uniformly accepted definition of MC CDMA.

Some define MC CDMA as a combination of orthogonal frequency division multiplexing(OFDM) and CDMA, where the number of carriers typically equals the processing gainwhile another definition is the parallel transmission of multiple narrowband DS wave-forms, where the number of such waveforms is typically much less than the processinggain. The reason for this is the likelihood that this latter model or, more precisely, a codedversion of it, as in [27], will be the first to be deployed in a wideband CDMA design. MCCDMA has the following characteristics;

- Similar to a conventional single carrier DS system, a multicarrier DS spread spectrumsystem is robust to multipath fading, as long as the total spread bandwidth exceeds thechannel coherence bandwidth.

- In a multicarrier DS system with M carriers, the entire bandwidth of the system isdivided into M equiwidth frequency bands, and thus each carrier frequency is modulatedby a spreading sequence with a chip rate which is 1/M times that of a single carrier system.Thus, a multicarrier system requires a lower speed, parallel type of signal processing, incontrast to a fast, serial type of signal processing in a single carrier rake receiver.

In essence, the use of this type of multicarrier design results in a tradeoff of the explicitpath diversity one would achieve by using a single carrier DS waveform (with the sametotal spread bandwidth) for the explicit frequency diversity achievable with the multicar-rier waveform. There can be many considerations in the choice of the parameter M. If westart out with a baseline system consisting of an MC system wherein the bandwidth ofeach of the carriers equals the coherence bandwidth of the channel, then increasing M tosome value, say, M1 > M, results in correlated fading among the carriers, the consequenceof which is an effective diversity order less than M1. Alternately, decreasing M to someother value, say, M2 < M, results in frequency selective fading on each carrier, and thus arake receiver on each sub carrier is required in order to achieve optimal performance.


Similarly, increasing M results in increased coding gain, but degreased processing gain persub carrier. Also, increasing M results in greater complexity at the transmitter, and, to theextent that the transmitter has a saturating power amplifier, increasing M results in moreintermodulation products among the carriers.

- Performance of MC CDMA: Consider a comparison of MC and single carrier CDMAcorresponding to the reverse (up) link, which is shown in figure 14 taken from [29].

FIGURE 14. BER versus Eb/ηo for K = 50, fj = f2, Wj = BWM, and α = 0.5

The figure shows the probability of error versus Eb/ηo for K=50, where K is the number ofsignals and Eb is the energy per bit. The curves are parameterized by JSR, where JSR isgiven by

JSR = interference power / signal power = nj * Wj / ( Eb / T )

where nj represents narrowband bandpass interference that is assumed to be Gaussian witha bandlimited double-sided power spectral density of nj/2, center frequency of fj, and abandwidth of Wj Hz. Such an interfering signal might, for example, represent a narrow-band waveform to be overlaid by the wideband MC CDMA network.

If JSR is small, the performance of a single carrier rake system and that of a multicarriersystem are almost the same. However, it is observed that the multicarrier system outper-forms the single carrier system if JSR is large.

- Forward (Down) Link Considerations: In the design of such links, because they can bemade synchronous with respect to the timing epochs of the symbols of the various signals,it is possible to employ orthogonal spreading sequences. Note however that if a multicar-


rier waveform is used such that M subcarriers span the spread bandwidth, then, relative toa wideband single carrier DS system, the number of orthogonal functions is reduced by afactor of M (assuming the system is designed just to maintain orthogonality within eachsubband). Hence there is the potential for significantly more multiple access interferenceon the forward link because of the use of multicarrier signaling, compared to the multipleaccess interference of a single carrier system, which only occurs because of multipath andout of cell interference. To help alleviate this situation, one can consider a hybrid scheme,which will end up having a tradeoff between the number of orthogonal waveforms and theamount of frequency diversity.

5.2.3 Multicode CDMA

Another possible change in the design of a CDMA network is to incorporate what isreferred to as multicode CDMA. Perhaps the key motivation for multicode CDMA is thatof increasing the information rate over a given spread bandwidth, an objective which isfundamental to future generation CDMA systems whereas to allow for the flexibility mul-tiple data rates can be an additional reason. If one attempts to increase the data rate simplyby decreasing the information symbol duration, while simultaneously keeping spreadbandwidth constant, the processing gain necessarily decreases. Indeed, such systems aretypically referred to as variable processing gain (VPG) systems. As an alternative to aVPG waveform, one can assign multiple spreading sequences to any given user. If thesespreading sequences are orthogonal, the self-interference caused by them is reduced.Clearly there are tradeoffs in the use of these two techniques to increase the data rate. Forhigh data rates, wherein processing gain can be very small, a reduction in the effectiveorder of diversity that the receiver can achieve is expected. This is because a reduced pro-cessing gain results in increased correlation among the taps of a rake receiver. Alternately,the multicode system suffers from increased interference, since each user now transmitsmultiple waveforms simultaneously, and each of those waveforms is affected by the multi-path on the channel.

The third-generation air interface standardization for the schemes based on CDMA seemsto focus on two main types of wideband CDMA: network asynchronous and synchronous.In network asynchronous schemes the base stations are not synchronized, while in net-work synchronous schemes the base stations are synchronized to each other within a fewmicroseconds. There are three network asynchronous CDMA proposals: WCDMA inETSI and in ARIB, and TTA II (Global CDMA II) wideband CDMA in Korea have simi-lar parameters. In addition, T1P1 in the United States has joined the development ofWCDMA. TR46.1 in the United States is also developing a wideband CDMA scheme,Wireless Multimedia & Messaging Services (WIMS), which has been recently harmo-nized with WCDMA. A network synchronous wideband CDMA scheme has been adoptedby TR45.5 (cdma2000) and is being considered by Korea (TTA I or Global CDMA I). Allschemes are geared towords the IMT-2000 radio transmission technology selection pro-cess in ITU-R TG8/1.

Several attempts have been made to harmonize the different wideband CDMA approachesin search of a unified global air interface. However, due to the evolution of current sys-tems and the strong commercial interests of their supporters, at the moment it seems that


there will be at least two wideband CDMA standards for third-generation: WCDMA andcdma2000.

REFERENCES

[1] Dahlman E., Gudmundson B., Nilsson M, and Skold J, UMTS/IMT-2000 Based onWideband CDMA, IEEE Communication Magazine, Sep. 1998 pp 70 - 80.

[2] Prasad Ramjee, Ojanpera Tero, An Overview of CDMA Evolution toward WidebandCDMA, IEEE Communication Surveys, 1998.

[3] Shannon, “A Mathematical Theory of Communication,” Bell System Technical Jour-

nal, Vol. 27, 1948, pp. 379 - 423 and 623 - 656.

[4] Scholtz, "The Evolution of Spread-Spectrum Multiple-Access Communications," in

Code Division Multiple Access Communications (S. G. Glisic and P. A. Leppanen, Eds.,

Kluwer Academic Publishers, 1995.

[5] "A conversation with Claude Shannon," IEEE Commun. Mag., Vol. 22, No. 5, May

1984, pp.123 - 126.

[6] Cooper and R. W. Nettleton, "A spread-spectrum techniques for high-capacity mobile

communications," IEEE Trans. Veh. Tech., Vol. 27, No. 4, Nov. 1978, pp. 264 - 275.

[7] Verdu, "Minimum Probability of Error for Asynchronous Gaussian Multiple Access,"

IEEE Trans. on IT., Vol. IT-32, No. 1, Jan. 1986, pp. 85 - 96.

[8] Ojanpera and R. Prasad, Wideband CDMA for Third Generation Mobile Communica-

tions, Artech House, Oct. 1998.

[9] Prasad, Universal Wireless Personal Communications, Artech House, Boston-London,

July 1998.

[10] Prasad, CDMA for Wireless Personal Communications, Artech House, Boston-Lon-

don, 1996.

[11] Garg, V. K., K. Smolik, and J. E. Wilkes, Applications of CDMA in Wireless/Per-

sonal Communications, Prentice-Hall, 1997.

[12] Viterbi, A. J., CDMA Principles of Spread-Spectrum Communications, Reading,

Mass.: Addison-Wesley Publishing Company, 1995.

[13] Gustafsson, M., K. Jamal, and E. Dahlman, "Compressed Mode Techniques for Inter-

frequency Measurements in a Wide-band DS-CDMA System," Proc. of PIMRC'97, Hels-

inki, Finland, Sep. 1997, pp. 23 - 35.

[14] TIA/EIA-95B, "Mobile Station-Base Station Compatibility Standard for Dual-Mode

Wideband Spread Spectrum Cellular Systems," Baseline Version, July 31, 1997.


[15] TIA/EIA-95A, "Mobile Station-Base Station Compatibility Standard for Dual-ModeWideband Spread Spectrum Cellular Systems" ,1995.

[16] ANSI J-STD-008, "Personal Station-Base Station Compatibility Requirements for 1.8to 2.0 GHz Code Division Multiple Access (CDMA) Personal Communications Systems",1995.

[17] TIA/EIA/IS-97, "Recommended Minimum Performance Standards for Base StationsSupporting Dual-Mode Wideband Spread Spectrum Cellular Mobile Stations," Dec. 1994.

[18] TIA/EIA/IS-98, "Recommended Minimum Performance Standards for Dual-ModeWideband Spread Spectrum Cellular Mobile Stations," Dec. 1994.

[19] Ross, A. H. M., and K. L. Gilhousen, "CDMA Technology and the IS-95 NorthAmerican Standard," ,in The Mobile Communications Handpaper, J. D. Gipson (Ed.),CRC Press, 1996, pp. 430 - 448.

[20] Milstein, Laurence B., ‘Wideband Code Division Multiple Access”, IEEE Journal on

Selected areas in Communication, Vol. 18, No: 8, August 2000, pp 1344 - 1354.

[21] T. Eng and L.B. Milstein, "Comparison of hybrid FDMA/CDMA systems in fre-

quency selective Rayleigh fading," IEEE J. Select Areas Communication, vol 12, pp. 938

- 951, June 1994.

[22] T. Eng and L.B. Milstein, "Partially coherent DS-SS performance in frequency selec-

tive multipath fading," IEEE Trans. Communication, vol. 45, pp. 110 - 118, Jan 1997.

[23] T. Eng, A. Chockalingam, and L.B. Milstein, "Capasities of FDMA/CDMA systems

in the presence of phase noise and multipath Rayleigh fading," IEEE Trans. Communica-

tion, vol. 46, pp. 997 - 999, Aug. 1998.

[24] Adachi, F., Sawahashi, M., and Suda, H., “Wideband DS-CDMA for Next Generation

Mobile Communications Systems”, IEEE Communications Magazine, Sep. 1998, pp 56 -

69.

[25] R.R. Rick and L.B. Milstein, "Parallel acquisition in mobile DS-CDMA systems,"

IEEE Trans. Communication, vol. 45, pp.1466 - 1476, Nov. 1997.

[26] R.R. Rick and L.B. Milstein, "Optimal decision strategies for acquisition of spread

spectrum signals in frequency selective fading channels," IEEE Trans. Communication,

vol. 46, pp. 686 - 694, May 1998.

[27] D.N. Rowitch and L.B. Milstein, "Coded multicarrier code division multiple access,"

in Proc IEEE Int. Symp. Inform. Theory, Sep. 1995, p. 23.

[28] S. Kondo and L.B. Milstein, "On the use of multicarrier direct sequence spread spec-

trum systems" in Proc. 1993 IEEE Military Commun. Conf., Oct. 1993, pp. 52 - 56.


[29] S. Kondo and L.B. Milstein, "Multicarrier DS CDMA systems in the presence of par-tial band interference," in Proc. IEEE Military Communication Conf., Oct 1994, pp.588 -599.

wcdma and cdma2000 - the radio interfaces for future...

Documents