30
CHAPTER 3
Wideband Code Division Multiple Access (WCDMA)
3.1 Introduction
The new requirements of the third generation systems are listed below:
Bit rates up to 2Mbps.
Variable bit rate to offer band width on demand.
Multiplexing of services with different quality requirements on a single connection,
e.g. speech, video and packet data.
Delay requirements from delay-sensitive real time traffic to flexible best effort
packet data.
Quality requirements from 10% frame error rate to 10-6 bit error rate.
Co-existence of second and third generation systems and inter system handovers for
coverage enhancements and load balancing.
Support of asymmetric uplink and downlink traffic, e.g. web browsing causes more
loading to downlink than to uplink.
High spectrum efficiency.
Co-existence of FDD and TDD modes.
Carrier spacing 5 MHz (normal)
Chip rate 3.84 Mcps
Frame length 10 ms (38400 chips)
Number of slots / frame 15
Number of chips / slot 2560 chips (normal)
Uplink SF 4 to 256
Downlink SF 4 to 512
Channel rate 7.5 Kbps to 960 Kbps
Table 3.1 Some parameters of WCDMA physical layer
31
3.2 Implementation of WCDMA
Fig 3.1 WCDMA system model
3.2.1 Channel izat ion Codes
There are certain restrictions as to which of the channelization codes can be
used for a transmission from a single source. Another physical channel may use a
certain code in WCDMA.
32
Specification Channelization code Scrambling code
Usage Uplink: Separation of
physical data (DPDCH) and
control channels (DPCCH)
from same terminal
Uplink: Separation of terminal
Downlink: Separation of
downlink connections to
different users within one
cell
Downlink: Separation of sectors
(cells)
Length Uplink: 4 – 256 chips (1.0
– 66.7 µs)
Uplink (1) 10ms = 38400 chips
Uplink (2) 66.7 µs=256 chips
(advanced BS receivers)
Downlink: 512 chips Downlink: 10ms = 38400 chips
Number of codes Number of codes under one
scrambling code =
spreading factor
Uplink: several millions
Downlink : 512
Code family OVSF Long 10ms = Gold code
Short – extended S(2) code family
Spreading Yes, increases Transmission
bandwidth
No, doesn’t affect transmission
bandwidth
Table 3.2: Functionality of the channelization and scrambling codes
the tree if no other physical channel to be transmitted using the same code tree is
using a code that is on an underlying branch, i.e. using a higher spreading factor code
generated from the intended spreading code to be used. Neither can a smaller
spreading factor code on the path to the root of the tree be used. The downlink
orthogonal codes within each base station are managed by the radio network
controller (RNC) in the network.
33
The functionality and characteristics of the scrambling and
channelization codes are summarized in Table 3.2 One code tree is used with
one scrambling code on top of the tree. This means that different terminals and
different base stations may operate their code trees totally independent of each
other; there is no need to coordinate the code tree resource usage between
different base stations or terminals.
3.2.2 Uplink spreading and modulation
3.2.2.1 Uplink modulation
In the uplink direction, there are basically two additional terminal-
oriented criteria that need to be taken into account in the definition of the
modulation and spreading methods. The uplink modulation should be designed so
that the terminal amplifier efficiency is maximized and the interference from the
terminal transmission is minimized.
With GSM operation, we are familiar with the occasional audible
interference with audio equipment that is not properly protected. The interference
from GSM has a frequency of 217 Hz, which is determined by the GSM frame
frequency. This interference falls into the band that can be heard by the human
ear. The continuous transmission achieved with an I-Q/code multiplexed control
channel is shown in Figure 3.4. Now, as the pilot and the power control signaling
are maintained on a separate continuous channel, no pulsed transmission
34
Fig 3.2 Parallel transmissions of DPDCH and DPCCH when data ispresent/absent (DTX)
occurs. The only pulse occurs when the data channel DPDCH is switched on and
off, but such switching seldom happens. The average interference to other users and
the cellular capacity remain the same as in the time- multiplexed solution. In
addition, the link level performance is the same in both schemes if the energy
allocated to the pilot and the power control signaling is the same.
For the best possible power amplifier efficiency, the terminal transmission
should have as low a peak-to-average (PAR) ratio as possible to allow the terminal
to operate with a minimal amplifier back-off requirement, mapping directly to
the amplifier power conversion efficiency, which in turn, is directly proportional
to the terminal talk time. With I – Q / code multiplexing, also called dual –
channel QPSK modulation, the power levels of the DPDCH and DPCCH are
typically different, especially as data rates increases, and would lead in extreme
cases to BPSK – type transmission when transmitting the branches independently.
This has been avoided by using a complex-valued scrambling operation after the
spreading with channelization codes.
The signal constellation of the I-Q/code multiplexing before complex
scrambling is shown in Figure 3.5. The same constellation is obtained after
35
descrambling in the receiver for the data detection G denotes the relative gain factor
between DPCCH and DPDCH branches.
.
Fig 3.3 Constellation of I-Q/code multiplexing before complex scrambling.
The transmission of two parallel channels, DPDCH and DPCCH, leads
to multicode transmission, which increases the peak-to-average power ratio
(crest factor). the transmitter power amplifier efficiency remains the same as
for normal balanced QPSK transmission in general. The complex scrambling
codes are formed in such a way that the rotations between consecutive chips
within one symbol period are limited to ±90°. The full 180° rotation can happen
only between consecutive symbols. This method further reduces the peak-to-
average ratio of the transmitted signal from the normal QPSK transmission.
The efficiency of the power amplifier remains constant, irrespective
of the power difference G between DPDCH and DPCCH. This can be
explained with the signal constellation for the I-Q/code multiplexed control
channel with complex spreading. In the middle constellation with G = 0.5, the
possible constellation points are only circles or only crosses during one symbol
36
period. Their constellation is the same as for rotated QPSK. Thus, the signal
envelope variations with complex spreading are
Fig 3.4. I-Q/code multiplexing with complex scrambling
Fig 3.5 Signal constellation for I-Q/code multiplexed control channel with
complex scrambling. G
very similar to QPSK transmission for all values of G. The I – Q /code
multiplexing solution with complex scrambling results in power amplifier
output back-off requirements that remain constant as a function of the power
difference between DPDCH and DPCCH.
37
3.2.2.2 Uplink Scrambling Codes
The transmissions from different sources are separated by the
scrambling codes. In the uplink direction, there are two alternatives: short and
long scrambling codes. The long codes with 25 degree generator polynomials are
truncated to the 10 ms frame length, resulting in 38 400 chips with 3.84 Mcps.
The short scrambling code length is 256 chips. The long scrambling codes are used
if the base station uses a Rake receiver (refer Appendix C). The Rake receiver is
described in next section. If advanced multiuser detectors or interference
cancellation receivers are used in the base station, short scrambling codes can be
used to make the implementation of the advanced receiver structures easier. Both
of the two scrambling code families contain millions of scrambling codes, thus,
in the uplink direction, code planning is not needed.
The short scrambling codes have been chosen from the extended S(2)
code family. The long codes are Gold codes. The complex-valued scrambling
sequence is formed in the case of short codes by combining two codes, and in the
case of long codes from a single sequence where the other sequence is the delayed
version of the first one.
The complex-valued scrambling code can be formed from two real-valued
codes C1 and C2 with the decimation principle as:
1 0 2 1(( )(2 ) )scramblingC C w jC k w (3.1)
Where 0 , 1 , 2 , . . .k
0 1
0 1{1 1} , {1 1}
w i th s e q u e n c e s w a n d w
w w (3.2)
38
The decimation factor with the second code is 2. This way of creating
the scrambling codes will reduce the zero crossings in the constellation and
will further reduce the amplitude variations in the modulation process.
3.2.2.3 Spreading and Modulation on Uplink Common Channels
The Random Access Channel (RACH) contains preambles that are sent
using the same scrambling code sequence as with the uplink transmission, the
difference being that only 4096 chips from the beginning of the code period are
needed and the modulation state transitions are limited in a different way. The
spreading and scrambling process on the RACH is QPSK-value, thus only one
sequence is used to spread and scramble both the in- phase and quadrature
branches. This has been chosen to reduce the complexity of the required
matched filter in the base station receivers for RACH reception.
The RACH message part spreading and modulation, including
scrambling, is identical to that for the dedicated channel. The codes available for
RACH scrambling use are transmitted on the BCH of each cell.
For the peak-to-average reduction, an additional rotation function is used
on the RACH preamble, given as:
( )4 2( ) ( ) 0,1, 2 , ..., 409 5
j k
b k a k e k
(3.3)
where a(K) is the binary preamble and b(K) is the resulting complex-
valued preamble with limited 90° phase transition between chips. The
autocorrelation properties are not affected by this operation.
39
The CPCH spreading and modulation are identical to those of the
RACH in order to maximize the commonality for both terminal and base
station implementation when supporting CPCH. RACH and CPCH
processes will be described in more detail in connection with the physical
layer procedures.
3 .3 Transmi t ter Characteri s t i c s
The pulse shaping method applied to the transmitted symbols is root-raised
cosine filtering with a roll-off factor of 0.22. The same roll-off is valid for both
the terminals and the base stations. There are a few other key RF parameters that are
introduced here and that have an essential impact on the implementation, as well
as on system behavior.
The nominal carrier spacing in WCDMA is 5 MHz but the carrier
frequency in WCDMA can be adjusted with a 200 kHz raster. The central
frequency of each WCDMA carrier is indicated with an accuracy of 200 kHz.
The target of this adjustment is to provide more flexibility for channel spacing
within the operator's band.
3.4 User Data Transmission
For user data transmission in second generation systems, such as the first
versions of GSM, typically only one service has been active at a time, either voice
or low-rate data. From the beginning, the technology base has required that the
physical layer implementation be defined to the last detail without real
flexibility. For example, puncturing patterns in GSM have been defined bit by
40
bit, whereas such a definition for all possible service combinations and data rates
is simply not possible for UTRA. Instead, algorithms for generating such
patterns are defined. Signal processing technology has also evolved greatly, thus
there is no longer a need to have items like puncturing on hardware as in the
early days of GSM hardware development. For the circuit switched (CS) traffic
(e.g. speech and video), a transmission dedicated channel needs to be used, while
for packet data, there are additional choices available, RACH and CPCH for the
uplink and FACH and DSCH for the downlink.
3.5 Multiple Input and Multiple Output on WCDMA
Mobile multimedia and high speed internet access are driving the development
of third generation (3G) wireless communication systems toward significantly higher
data rates than those achieved by their second generation (2G) predecessors. To
provide this the system must support a large number of concurrent users. Using
DSCDMA, the UMTS standard allows different data rates to be allocated across the
active users by appropriate selection of spreading factors, with a maximum data rate
of 2 Mb/s. However, assuming QPSK modulation, the total uncoded data rate of the
system is limited to 8 Mb/s. A detailed analysis [12] has shown that these limits are
almost never consistently achieved in DS – CDMA systems using conventional
RAKE receivers due to multiple access interference in multipath environments.
The application of multiple antennas is expected to increase system capacity
through different mechanism. The most straight forward approach is to exploit spatial
diversity with multiple antennas at the transmitter and/or receiver. Different
techniques are used to mitigate the effects of fading and spatially non – white
41
interference (in – cell and out – of – cell multiple access interference) thereby
increasing the received SINR. As such, one possible method of spatial diversity is
already defined within the UMTS standard [13]. While these methods allow the
system to Improve the transmission data rates, neither transmitter nor receiver
diversity techniques significantly increases the system capacity. In addition, some of
them are disadvantaged by their required knowledge of the channel at the transmitter.
In 1996 Foschini pointed out [14] that the use of multiple antenna at the
transmitter and at the receiver would allow for a significant increase in system
capacity through the application of spatial multiplexing. It is shown that the capacity
of such a system with M transmitter and N receiver antenna is:
2log det HM M
SNRC I HH
M
(3.4)
where H describes the M×N channel between the M transmit and N receive antennas
(the matrix elements are assumed to be complex and unit variance variables). Under
idealized conditions (e.g., orthogonal channels and M = N) this spatial multiplexing
results in an M fold capacity gain [15].
A rich scattering environment is known to favor spatial multiplexing. We refer
to this form of transmission over multiple antenna systems as BLAST (Bell Labs
Layered space time). So far the concept has been analyzed, demonstrated and verified
for narrow band system in numerous publications [16] – [21]. A study that compares
capacity and BER performance of space time coding, beam forming and BLAST in a
UMTS TDD scenario has been presented in [22] and [23]. In summary, it is
concluded that in typical indoor environments and when no reliable channel state
42
information is available at the transmitter (as in the UMTS – FDD downlink) spatial
multiplexing is the method best suited for optimally exploiting the channel capacity.
Further, some theoretical and practical performance aspects will be
highlighted. Next section provides a description of the implementation of the main
components in the system. Furthermore, the testing of the system and corresponding
results are presented and prototype implementation is described. Next section
presents indoor over – the – air BER and capacity measurement results.
3.5.1 MIMO Extended UMTS – FDD downlink.
The design is based on the DS – CDMA principle and uses the UMTS - FDD
chip – rate of 3.84MHz [2, 13]. It’s essential components are the dedicated physical
channel (DPCH), a common pilot channel (CPICH) for the pilot assisted channel
estimation and the primary and secondary synchronization channels (PSCH, SSCH).
The common pilot channels, as well as the user channels, are separated by utilizing
orthogonal variable rate channelization codes, which are derived from Hadamard
matrix (OVSF – Codes). The CPICH uses a length 256 OVSF code, while the length
of the codes for the DPCH is variable to support raw data rates between 31.25kb/s and
2Mb/s in the single antenna case and 125kb/s to 8Mb/s in the proposed 4 × 4 antenna
configuration. After channelization, the channel are scrambled with the same
truncated Gold sequence. Adjacent base stations reside in the same band and are
separated through the use of different Gold sequences. The PSCH and the SSCH are
added after the scrambling and are not orthogonal to the other codes. They are both
required to achieve initial synchronization and to identify the Gold sequence used by
the targeted base station [13].
43
The extensions of this basic system to support M transmit and N receive
antenna for spatial multiplexing is straight forward. After channel coding, the
incoming serial data stream of each user in this mode is first demultiplexed into M
parallel streams. Each of the stream uses the same channelization code and is
scrambled with the same Gold sequence before it is transmitted through one of the M
transmit antennas. The total transmitted power is begin conserved by sharing it
equally among the antennas and no additional user codes are required. It should also
be noted that other users, which do not use this scheme, are not affected, as the
codes of the MIMO users are still orthogonal to the other channelization codes in
the system. However, while the legacy system requires only one pilot channel for
the estimation of the 1×1 channel at the mobile, this is not sufficient for the pilot
assisted estimation of the M×N channels of the MIMO systems. Therefore, a
separate pilot channel, again using a length 256 OVSF code is assigned to each
of the transmit antennas.
The synchronization channels utilize only one of the transmit antennas,
as their replication would lead to an inherent beam forming, which is not
intended. The transmitter of the realized system with M = 4 transmit antennas.
Only two users are depicted in this diagram for readability. Many details of the
implemented system are presented in [14] – [18]. It has been shown that large
diversity gains are desired to reduce the effect of fading.
44
Fig 3.6 Basic block diagram for MIMO communication
3 . 5 . 2 Signal Model and MIMO rake receiver.
It has been shown that large diversity gains are desired to reduce the effect of
fading. It is therefore concluded that multipath propagation should be exploited
whenever it exits [19]. In general, adding multipath components does not
automatically lead to better performance (see the work in [11, 20, 21] ). In addition,
for practical receiver designs multipath is not necessarily a desirable condition as it
can also lead to serious multiple access and self interference, reducing the effective
system capacity [22]. The severity of the interference depends strongly on the type of
receiver. While a MIMO equalizer (temporal and spatial) approach is a method to
avoid interference and still exploit diversity, its complexity is often prohibitive,
especially in high mobility environments where fast update rates are necessary. A
MIMO receiver offers the ability to exploit the advantages of the multipath
propagation at a reasonable implementation cost. However, this is done at the cost of
additional interference.
It has been shown that the receive diversity scheme is not suitable for the
downlink, as it is difficult and inconvenient to install multiple antennas on handsets
45
[106]. The multiple antenna burden is preferably placed at the base station. This is
called transmit diversity. Transmit diversity has gained a lot of attraction and research
in last five years. In transmit diversity the same signal is transmitted from all
antennas. If same signal is transmitted from all the antennas, at the receiver the copies
of this signal add incoherently, and no diversity gain can be achieved. Thus in order to
diversity transmission to work, one must find a transmission scheme where replicas of
the signal combine coherently at the receiver. One of the simplest and most attractive
transmit diversity schemes was proposed by Alamouti [10,109].
S1 S2
G2 = (3.5)
-S*2 S*
1
Here the rows denote time instances and columns denote transmit antennas. At time t1
= 1, S1 and S2 will be transmitted from antennas 1 and 2 respectively, and at time t2, -
S2 *and S2 will be transmitted from antennas 1 and 2 respectively. One can see that
two symbols are transmitted over two time intervals. Hence the code is full rate.
Assuming the fading coefficients as h1 & h2 for the transmitting antenna 1 and 2 is
assumed to be constant over nT = 2 consecutive time slots.
h1 = h1 (T = 1) = h1 (T=2) (3.6)
h2 = h2 (T=1) = h2 (T=2) (3.7)
Hence the received signal is
Y1 = h1x1 + h2x2 + n1 (3.8)
Y2 = -h1x*2 + h2x
*1 + n2 (3.9)
46
We observe from the expression that by using two transmit and one receive
antenna, the transmitted signals are effectively multiplied by h12 + h2
2 . Hence if one
of the paths is in fade, the other may still represent the signal with reliability. In fact,
the use of orthogonal STBC changes the probability distribution of the channel to
distribution with lower variance. It can also be observed that STBC in MIMO
channels can be represented as equivalent SISO channel. Tarokh et al[12] extended
Alamouti's 2-transmit diversity scheme to more than two antennas. This new
generalized space-time signaling scheme is known as Space Time Block Codes
(STBC). Space Time Block Codes derive their name from the fact that the encoding is
done in both space and time, a simple a matrix defines their encoder. A space time
block code is defined by the relationship between the k-tuple signal x and the set of
signals to be transmitted from nT antenna over p time periods. Such a relation is given
by p x nT transmission matrix.
S11 S12………….S1nT
G = S21 S22………….S2nT (3.10)
.... … …….. ….
SP1 SP2………..SpnT
Where sij are functions of k-tuple input sequence x1;x2;:::;xk and their
complex conjugates. At time slot 1, sij is transmitted from antenna j. Since k
information bits are transmitted over p time interval, the rate of the code is defined as
R= k/p. At the receiver we can use arbitrary number of receive antennas. The design
does not depend on the number of receive antennas nR. At the receiver, the nR
47
receive antennas use maximum likelihood (ML) decoding. In orthogonal STBC and
ML decoding is equivalent to Maximum Ratio Combining (MRC). Assuming perfect
channel side-information (SI) the decoder at antenna j maximizes. Since the block
coding requires only linear processing at the receiver, the decoding can be done
efficiently and quickly. Space Time Block Codes can be constructed for any type of
signal constellation and provide full diversity.
Exploiting multiple antennas at the transmitter and the receiver in wireless
communication systems has recently been proven to provide substantial benefits in
both increasing system capacity and improving its immunity to deep fading in the
channel [1, 2]. To take advantage of these benefits, special space-time coding
techniques are required. A pioneering work in the area of space-time coding for
multiple-input multiple-output (MIMO) wireless channels has been done by Tarokh et
al. in [3], in which two code design criteria have been proposed for flat fading
channels with coherent receivers, and high-performance space-time trellis codes have
been designed. However, these codes suffer from rather high decoding complexity. In
the same year, Alamouti proposed his celebrated space-time block coding (STBC)
scheme for two transmit and multiple receive antennas [4]. The maximum likelihood
(ML) decoder for Alamouti’s code has very low complexity. Inspired by this work,
Tarokh et al. generalized Alamouti’s codes to multiple transmit antennas using the
theories of orthogonal and amicable designs [5]. The codes developed in [5] are
known as orthogonal STBCs (OSTBCs) and are able to provide high performance at
very low decoding complexity. Some other designs of OSTBCs have also been
recently developed in [6–8].
48
OSTBC theory can be divided into two almost separate parts. The first part
deals with how OSTBCs can be designed. This aspect of the OSTBC theory is beyond
the scope of our chapter, and interested readers are referred to the relevant literature
[5–9]. The second part of this theory deals with the properties of these codes and
constitutes the focus of our chapter. Interestingly, the main properties of OSTBCs can
be derived from their simple definition proposed in [5] and are almost independent of
the way these codes have been constructed. In this chapter, we provide a background
of the theory of OSTBCs and review recent results that are based on their
constellation space invariance property and similarities between OSTBC models and
signal models used in array processing.
49
3.6 Algorithm
Fig 3.7 Flow diagram
3.7 Results
The above shown algorithm is executed using MATLAB and outputs are
tabulated. The graph has been plotted between power in (dB) and frequency (Hz) on
50
both the channels (AWGN & Rayleigh). Fig 3.9 represents baseband signaling of
power spectrum. Fig 3.10 represents the modulated signal at the channels plotted
against Amplitude and Time. Fig 3.11 represents the modulated signal spectrum. Fig
3.12 represents Output signal spectrum. Fig 3.13 represents Constellation output for
the signal before and after the channel. Both theoretical and simulated estimation of
AWGN and Rayleigh noise channel are shown in the Fig 3.14. The performance
analysis of 2 x 2 MIMO with different users is shown in the Fig 3.15.
Fig 3.8. Output waveform obtained from various levels of blocks
51
Output signal observed along with various parameters in accordance with same input
Fig 3.9. Baseband signal power spectrum for AWGN (Left) and Rayleigh (Right) in
both the graph frequency (Hz) differs
Fig 3.10. Modulated signal at the channel for AWGN (Left) and Rayleigh (right)
Fig 3.11. Modulated signal spectrum for AWGN (Left) and Rayleigh (Right)
52
Fig 3.12. Output signal spectrum for the received signal
Fig 3.13. Constellation output for the signal before and after the channel
53
3.14 BER analysis of AWGN and Rayleigh channel for both theory and simulated
Fig 3.15 Performance Analysis of 2×2 Mimo Antenna With Different Users
3.8 Summary
The MATLAB® simulation test bed was written and used to extract the
main characteristics of the target application. Once broad guidelines of the design
process were available, a synthesizable VHDL description of the architecture was
written. The design of the architecture was further optimized through iterative post
synthesize simulations and redesign.