Multicarrier CDMA Techniques

Multicarrier CDMA Techniques
As discussed in the previous section, a multi-carrier scheme is used to implement an OC code–based
CDMA system, if the FDM technique is used to send different element codes.
There is big difference between an OC code–based CDMA system implemented by a multicarrier
scheme and a general multi-carrier CDMA system, which is the focal point in this section. An
OC code–based CDMA system uses a multi-carrier architecture to send different element codes of
a complementary code assigned to a specific user. Thus, different frequency channels convey totally
different information, each piece of which is indispensable to the successful reconstruction of ACFs or
CCFs at a receiver. In this way, there is no frequency diversity in an OC code–based CDMA system.
On the other hand, a conventional multi-carrier CDMA system uses different carrier frequencies to
multiplex the same input wideband data stream, such that each frequency carries only a narrowband
bit stream, whose transmission rate is only 1
M -th of that of the input wideband stream if M subcarriers
are used. Each narrowband subcarrier channel is less vulnerable to frequency-selective fading if the
coherent bandwidth of the channel is wider than the bandwidth occupied by each subcarrier channel.
Even if some subcarrier channels unfortunately fail into the nulls of deep frequency fades, the joint
use of interleaving and error-correction coding can effectively recover the information corrupted by
the deep fade nulls in those subcarrier channels.
Therefore, multicarrier modulation (MCM) is the principle of transmitting data by dividing the
input wideband stream into several parallel narrowband bit streams, each of which has a much  lower bit rate, and using these substreams to modulate different carriers. On the basis of this
principle, several derivative MC-CDMA schemes have been introduced and they have been studied
extensively in [571–584]. On the basis of the division on duplication or multiplex schemes,
there are two MC-CDMA forms, that is, (1) duplicated MC-CDMA (2) multiplexed MC-CDMA;
on the basis of which domain the DS spreading takes place, there are also two alternative schemes,
which are (1) time-spreading MC-CDMA, and (2) frequency-spreading MC-CDMA. Therefore, by
mixing all of them in various combinations, we will have four different MC-CDMA systems as
follows:
• Duplicated time-spreading MC-CDMA;
• Duplicated frequency-spreading MC-CDMA;
• Multiplexed time-spreading MC-CDMA;
• Multiplexed frequency-spreading MC-CDMA.
The term duplicated or multiplexed means that the data streams in different subcarrier branches in a
transmitter are the “duplicated” or the “multiplexed” version of the input data stream, respectively.
We give a brief introduction to them in this section.
7.4.1 Duplicated Time-Spreading MC-CDMA
A duplicated time-spreading MC-CDMA scheme, in which M subcarriers are used, is illustrated in
Figure 7.8. The input data stream is duplicated in M subcarrier branches, which undergo spreading
modulation using the same signature code assigned to the k-th user. Only one transmitter for the
k-th user is shown in the figure for simplicity; while the signal of interest to the receiver is the k-th transmission. Time-domain spreading takes place in this scheme, and thus the input data stream will
be converted into a wideband subcarrier-modulated signal before it is sent into the channel.
Obviously, this scheme provides M duplicated SS signals in all M subcarriers. In other words,
the signals conveyed in M different subcarrier channels are exactly the same replicas of the original
input data. Therefore, there is a great deal of redundance in the transmission, which can be explored
to offer sufficient frequency diversity for signal detection. In this way, the scheme is very robust
in terms of its immunity against frequency selectivity of the channel. Of course, the price paid for
this strong immunity is the consumption of a great amount of bandwidth resource in sending all
M identical data streams in different subcarriers. Thus, the bandwidth efficiency of the duplicated
time-spreading MC-CDMA scheme is relatively low. For this reason, this scheme is not a popular
option of MC-CDMA systems.
7.4.2 Duplicated Frequency-Spreading MC-CDMA
Figure 7.9 depicts a conceptual block diagram of a duplicated frequency-spreading MC-CDMA, where
M subcarriers, or {f1, f2, . . . , fM}, are used. Only the k-th transmitter is shown in the figure and
the receiver is dedicated for the k-th user’s transmission. The k-th user is assigned a signature code
{ck1
, ck2
, . . . , ck
M}.
It is noted that DS spreading takes place in the frequency domain, and thus a matched filtering
is applied to the receiver at the outputs from all M low-pass filters to yield a decision variable. All
subcarriers are orthogonal with each other, with each overlapped with its neighboring subchannels
by half, to improve the overall bandwidth efficiency.
The multi-carrier modulated signals are not spread in the time domain in this scheme. The PG
of this system will be achieved in the frequency domain, where MAI will be rejected only at the
last stage of the receiver, that is, the frequency-domain matched filter. In this duplicated frequencyspreading
MC-CDMA scheme, the PG value is exactly equal to the number of subcarriers or M,
which becomes a very important system parameter of the scheme.
In the duplicated frequency-spreading MC-CDMA scheme, all subcarriers convey the same data,
but are spread modulated by different chips. Therefore, this scheme provides redundancy in the
frequency domain, but it will not help improve the system’s robustness against the frequency-selective
fading caused by the multipath propagation effect. The reason is because spreading over the different subcarrier frequencies, {f1, f2, . . . , fM}, requires that each subcarrier channel should give the identical
gain; otherwise it will cause a serious problem in the reconstruction of ideal ACFs and CCFs at the
matched filter in a receiver, thus introducing serious MAI.
Figure 7.9 does not include the interleaving and error-correction coding blocks for simplicity of
illustration. However, a practical duplicated frequency-spreading MC-CDMA system should always
work with them to achieve a better performance against MI.
7.4.3 Multiplexed Time-Spreading MC-CDMA
Figure 7.10 illustrates a conceptual block diagram of a multiplexed time-spreading MC-CDMA system,
where the k-th transmitter and a receiver tuned to the k-th transmission are shown. It is seen
from the figure that the input data stream {ak
1, ak
2, . . . , ak
M} with a bit rate of 1
T is first multiplexed
into M substreams, each of which has a rate of 1
MT and will be spread modulated by the same signature
code ck (t) assigned to the k-th user, followed by carrier modulation using subcarrier fi, where
i = 1, 2, . . . , M. The spectra of all subcarriers overlap with one another with a spectral offset of 1
2MT ,
thus forming an orthogonal multi-carrier CDMA signaling.
The received signal in this MC-CDMA scheme will be fed into a receiver. First, it undergoes
subcarrier demodulation using all M different subcarriers, followed by spreading demodulation using
the signature code ck (t), which is assigned to the k-th transmitter. The use of this signature code in
the receiver allows CDMA in this MC-CDMA system. The signal in each branch of the MC-CDMA
receiver will go through low-pass filtering, the decision device, and then the de-multiplexing unit to
yield a recovered wideband data stream, {ˆak
1 , ˆak
2, . . . , ˆak
M}.
In the multiplexed time-spreading MC-CDMA scheme, each subcarrier branch carries different
information due to the use of the multiplexing unit or the serial-to-parallel unit at the transmitter.
Thus, no redundance exists in different subcarriers. However, each subcarrier will convey only a
narrowband signal and it will only experience a flat fading, instead of a frequency-selective fading.
With the help of interleaving cum error-correction algorithms, this MC-CDMA scheme can work
satisfactorily in a channel with MI. In addition, a possible multipath diversity can be achieved if a RAKE receiver can be applied to replace a conventional matched filter, as used in Figure 7.10, that
is, the despreading unit and low-pass filter (LPF) in each subcarrier branch in the receiver.
7.4.4 Multiplexed Frequency-Spreading MC-CDMA
Finally, we would like to introduce the last MC-CDMA scheme called Multiplexed frequencyspreading
MC-CDMA, whose block diagram is shown in Figure 7.11.
In this MC-CDMA scheme, the wideband input data stream is multiplexed into M narrowband
substreams, each of which should be carrier modulated by M different subcarriers {f1, f2, . . . , fM}.
Frequency-domain spreading takes place across all subcarriers in this scheme. Therefore, MAI suppression
is ensured using some spreading codes with satisfactory CCF between any pair of them.
Thus, there is no redundance in the frequency domain and different subcarriers deliver totally different
information. Also, because of the use of frequency-domain spreading, the transmitting signal
will not be spread in time. Obviously, this scheme will offer the highest bandwidth efficiency of all
the four MC-CDMA schemes discussed in this section. However, this scheme does not provide any
frequency diversity and time diversity, and its performance cannot be comparable to any of the other
MC-CDMA schemes.
Having introduced the four different MC-CDMA schemes, namely, duplicated time-spreading
MC-CDMA, duplicated frequency-spreading MC-CDMA, multiplexed time-spreading MC-CDMA,
and multiplexed frequency-spreading MC-CDMA, we can compare them in terms of different performance
parameters, as given in Table 7.4.