Академический Документы
Профессиональный Документы
Культура Документы
CHAPTER-2
SPREADING CODES
This chapter deals with detailed description of spreading codes that include
are orthogonal to each other, and they are generated by using the
--- 2.1
where HN contains the same but inverted elements of HN . The seed matrix is
--- 2.2
Therefore, to derive a set of four orthogonal Walsh sequences w0, w1, w2, and w3, we only
--- 2.3
The four orthogonal sequences in this Walsh code set are taken from the rows of the
--- 2.4
For DS-SS multiple access[8,9], three conditions that must be met by a set of orthogonal
2. Each sequence in the set has an equal number of 1s and -1s, or the number of 1s differs
By changing the 0s to -1s in each of the four sequences above, that is,
--- 2.5
The calculation of cross products and dot products can be facilitated. In general, the 0th
Walsh sequence consists of all -1s and thus cannot be used for channelization. In the IS-
Equation (2.1) can be recursively used to generate Hadamard matrices of higher orders in
order to obtain larger sets of orthogonal sequences. For example, 8 orthogonal sequences,
length 16, can be obtained by generating H16. The IS-95 forward link uses a set of 64
Chapter-2 Spreading Codes 16
orthogonal Walsh sequences[8], thus the physical limitation on the number of channels on
the forward link is 63 because in an IS-95 system, w0 is not used to transmit any base-
band information.
--- 2.6
--- 2.7
Chapter-2 Spreading Codes 17
The following example illustrates how Walsh codes can be used for DS-SS multiple
access. Suppose that there are three different users, and each user wishes to send a
---2.8
--- 2.9
Each message is spread by its assigned Walsh code. Note that the chip rate of the Walsh
code is four times the bit rate of the message, contributing to a processing gain of 4. For
message one:
Chapter-2 Spreading Codes 18
Note that m1(t)w1(t) is the spread-spectrum signal of the first message. Similarly, for
message two:
The spread-spectrum signals for all three messages, m1(t)w1(t), m2(t)w2(t), and m3(t)w3(t),
--- 2.10
C(t) is the composite signal that is transmitted in the single RF band. If there are
negligible errors during the transmission process, the receiver intercepts C(t). In order to
separate out the original messages m1(t), m2(t), and m3(t) from the composite signal C(t),
the receiver multiplies C(t) by the assigned Walsh code for each message:
Then the receiver integrates, or adds up, all the values over each bit period. The functions
A “decision threshold” looks at the integrated functions M1(t), M2(t), and M3(t). The
--- 2.11
2.3 M-SEQUENCES
A simple shift register generator (SSRG) has all the feedback signals returned to a single
input of a shift register (delay line). The SSRG is linear if the feedback function can be
The feedback function f( x1,x2,…..xn ) is a modulo-2 sum of the contents xi of the shift
register cells with ci being the feedback connection coefficients (ci=0 is open, ci=1 is
connect). An SSRG with L flip-flops produces sequences that depend upon register length
L, feedback tap connections and initial conditions. When the period (length) of the
number of taps. If an L-stage SSRG has feedback taps on stages L, k,m and has
sequence……ai, ai+1, ai+2,…… then the reverse SSRG has feedback taps on L, L-k, L-
[5,2]=-1 -1 -1 -1 1 -1 1 -1 1 1 1 -1 1 1 -1 -1 -1 1 1 1 1 1 -1 -1 1 1 -1 1 -1 -1 1
[5,3]=-1 -1 -1 -1 1 -1 -1 1 -1 1 1 -1 -1 1 1 1 1 1 -1 -1 -1 1 1 -1 1 1 1 -1 1 -1 1
Chapter-2 Spreading Codes 22
2.4 PROPERTIES[10]
1 Balance:
For an m-sequence there is one more “one” then, “zero” in a full m-period of the
sequence. Since all states but the ‘all zero’ state are reached in an m-sequence, there must
2 Run-Length distribution:
For every m-sequence period, half the runs (of all 1’s or all 0’s) have length 1, one fourth
have length 2, one eight have length 3, etc. for each of the runs there are equally many
3 Autocorrelation:
The autocorrelation function of the m-sequence is -1 for all values of the chip phase shift
τ except for the [-1, +1] chip phase shift area in which correlation varies linearly from the
4 Cross-correlation:
The cross-correlation is the measure of agreement between two different codes.
numbers of transmitters, using different codes, are to share a common frequency band
5 Security:
The m-sequences codes are linear, and thus not usable to secure system. The linear codes
are easily decipherable once a short sequential set of chips (2L+1) from the sequence is
known.( The overall system could still be secure if the information itself where encoded
environment (CDMA) needs a set of codes with the same length and with good cross-
correlation properties.
Gold code sequences are useful because a large number of codes (with the same length
and with controlled cross-correlation) can be generated, although they require only one
Gold codes are product codes achieved by the exclusively OR-ing (modulo-2 adding) of
two maximum length sequences with the same length (factor codes). The code sequences
are added chip by chip by synchronous clocking. Because the m-sequences are of the
same length as the two base codes which are added together, but are non maximal (so the
auto-correlation function will be worse then that of m-sequences). Every change in phase
position between the two generated m-sequences causes a new sequence to be generated.
Chapter-2 Spreading Codes 24
Any two register gold code generator of length L can generate 2L-1 sequences (length
2L-1) plus the two base m-sequences, giving a total of 2L+1 sequences.
In addition to their advantage in generating large number of codes the gold codes may be
chosen so that over a set of codes available for a given generator the autocorrelation and
the cross-correlation between the codes is uniform and bounded. When spatially selected
m-sequences are used, the generated gold codes have a three valued cross-correlation.
The binary sample space for the design of the proposed new orthogonal codes[3,4] is
limited only to the zero mean (except DC sequence) and linear phase codes in this study.
Hence, for the case of 8-bit (length 8) codes, binary sample space consists of 22 unique
codes, and for 16-bit codes there are only 326 candidate codes, and similarly, for 32-bit
codes there are about 38,000 potential codes[3,4]. Brute force search is performed in order
Walsh codes perform poorly in AWGN channels for asynchronous communications when
the circular shift of the second code (following code in a stream of codes) matches with
the first code or the complement of the first code for any particular chip delay (high
cross-correlation). Note that the decimal values of all n-bit (size n) Walsh codes in a set
are multiples of 2(n/2) - 1 or 2(n/2) +1. For example, the codes of size 16 Walsh set are
multiples of 255 or 257 in decimals. In our orthogonal code design and brute force
search[3,4], we avoided such strict code conditions and the number of codes in the set that
Orthogonal code sets with linear phase are iteratively searched and selected from the
binary sample space for the given dimension. In any n-bit sample space (n-dimensional
Chapter-2 Spreading Codes 26
space), n-1 orthogonal sets can be formed and the DC code is added in order to make it a
complete binary set. Using this brute force search method, a number of independent
binary code sets are obtained from the sample space. Note that the degree of freedom is
very limited when the dimension of the space is low, i.e. short codes or small values of n.
other sizes like 7, 11, 15 and many more. Near orthogonal refers to the code sets having
their in-phase cross correlation value of -1. For near orthogonal code set design, instead
of limiting the code space to linear phase codes, entire binary space is considered. The
spectrum radio applications like dense sensor networks with low power and low
complexity constraints. Similar limitations also exist in the Gold family of binary codes
and the size and the number of available codes is very strict and limited due to their
nature. Similarly, a typical set of size 32 binary sequences for the proposed family along
with the Walsh set of the same size are also displayed in Table 4.2. Note that except the
first sequence, which is DC, none of the basis functions of the new orthogonal binary set
is common with the Walsh set for 32-bit codes. There are several other 32 dimensional
Aperiodic cross-correlation sequence between a typical pair of codes (2-user case with
C xy (m) for positive values of m is displayed in Figure 4.4 for the three binary families
under consideration. It is observed from the figure that the Gold and the proposed
Even (periodic) cross correlation is the cyclic correlation between the code vectors and it
is defined as
Traditional ways of separating signals in time (i.e., time division multiple access,
(TDMA)), or in frequency (i.e., FDMA) are relatively simple ways of making sure that
the signals are orthogonal and noninterfering. However, in CDMA, different users occupy
the same bandwidth at the same time, but are separated from each other via the use of a
--- 2.17
In discrete time, the two sequences x and y are orthogonal if their cross-product Rxy(0) is
--- 2.18
where
--- 2.19
--- 2.20
Note that T denotes the transpose of the column vector, which is another representation of
a sequence of numbers. For example, the following two sequences or codes, x and y, are
orthogonal:
--- 2.21
--- 2.22
--- 2.23
In order for the set of codes to be used in a multiple access scheme, we need two
additional properties. In addition to the zero cross-correlation property, each code in the
set of orthogonal codes must have an equal number of 1s and –1s[11]. This second property
Chapter-2 Spreading Codes 29
gives that particular code the pseudorandom nature. The third property is that the dot
product of each code scaled by the order of the code must equal to 1. The order of the
code is effectively the length of the code, and the dot product is defined as a scalar
obtained by multiplying the sequence by itself and summing the individual terms; that is,
--- 2.24
The two orthogonal codes in the previous example also satisfy the second and the third
conditions. Both x and y have an equal number of 1s and –1s, and the scaled dot products
are
---2.25
---2.26
The properties of the set of orthogonal codes to be used in DS-SS multiple access:
2. Each sequence in the set has an equal number of 1s and –1s, or the number of 1s differs
Figure 1.1 shows the waveforms and spectrums for the two messages m1(t) and m2(t), the
two orthogonal codes c1(t) and c2(t), and the two spread messages m1(t)c1(t) and m2(t)c2(t).
While we do not go into the details of calculating the spectrums of these time waveforms,
it suffices for our purposes to state that the bandwidth of a random digital waveform is
limited to 1/T, where T is the bit interval of the random digital waveform. We further
Chapter-2 Spreading Codes 30
make the distinction between Tb and Tc , where Tb is the bit interval (in seconds) of the
message and Tc is the chip interval of the running orthogonal code. In this example, the
chip rate (1/Tc ) of the orthogonal code is running at four times the bit rate (1/Tb ).
expansion factor is sometimes called the processing gain or (W/R), where W is the final
bandwidth of the spread message and R is the bandwidth of the baseband message. Note
Note that after spreading by the orthogonal codes, the spread messages m1(t)c1(t) and
Figure 2.5 shows the waveforms at different points of the receiver. The signal at point A
is the result of the summation of the two spread messages. The spectrum at A now
contains two separate signals. In order to recover the two separate messages from the
composite spectrum, the signal at A is multiplied by the two respective orthogonal codes
to obtain B1 and B2. This example only serves to illustrate the principle of DS-SS multiple
access. It has been explained that, using DS-SS techniques, separate messages can be sent
through the same channel in the same frequency band at the same time, and the messages
can be successfully recovered at the receiver. However, there are many real-world
Figure 2.5 Time waveforms and frequency spectra for the base-band messages m1( t)
and m2( t), orthogonal codes c1( t) and c2( t), and spread messages m1( t) c1( t) and
m2( t) c2( t).
Chapter-2 Spreading Codes 33
Figure 2.6 Time waveforms and frequency spectra for the signals at different points
of the receiver.
Chapter-2 Spreading Codes 34
There are two problems: the near-far problem, and the partial
correlation problem[11-13].
are closer to the base station than others. The result is that powers
received from those users that are close by are higher than powers
received from users that are farther away. Because all users are
transmitting in the same band, the higher received powers from users
received at the base station are the same for all users.
band. Even when the transmitters are synchronized, there is still the
problem of propagation delay, which is inherent in a mobile channel. For example, the
two codes mentioned previously are orthogonal when they are perfectly aligned:
channel, then
Chapter-2 Spreading Codes 35
If the codes are not orthogonal due to synchronization or channel impairment, then
multiple access messages in the same band can no longer be separated from one another
via code-orthogonality. The results are correlation crosstalk and mutual interference. In
--- 2.27
--- 2.28
Therefore, simple orthogonality between two aligned codes is not enough the above two
partial correlations must also be zero, or at least small, for any value of t likely