2
Optical Coding Schemes
Since the mid-1980s, there have been steady developments in coding techniques and
enabling techn o logies in the field of coding-based optical systems and networks,
exemplified by Optical Code-Division Multiple Access (O-CDMA) and multiplex-
ing [1–34]. With the advances in optical hardware tewachnology, the large bandwidth
expansion required by optical codes can now be accommodated and multiple-user
O-CDMA experimental testbeds opera tin g at 10 Gbits/s have been demonstrated
[34–37]. For instance, there are two main categories of optical coding schemes—
synchronous and asynchronous—depending on whether synchronization (usually in
time or wavelength) is needed among optical codewords. Whilebothkindsofcod-
ing schemes have their pros and cons, asynchronous coding, ingeneral,allowssi-
multaneous users to access the same optical transmission medium independently
with no wait time, sched uling, or coordinatio n. Compared to conventional asyn-
chronous schemes, such as Carrier-Sense Multiple Access with Collision Detection
(CSMA/CD) [38], asynchronous coding makes more efficient useofthetransmis-
sion medium because a user does not need to wait for the medium to become idle
before gaining access. The number of collisions among the signals of simultane-
ous users increases and access delay gets worse when conventional asynchronous
schemes are used in a high-speed or heavy traffic-load environment. Asynchronous
coding is, however, m ore suitable for this kind of system environment. Furthermore,
because the traffic in a local area network is typically bursty, asynchronous multiple-
access schemes, such as incoh er en t O-CDMA, are more efficientthansynchronous
multiple-access schemes, exemplified by Optical Time-Division Multiple Access (O-
TDMA), in which a fixed portion (i.e., time) of the m edium is dedicated to a partic-
ular user.
In addition to O-CDMA and O-TDMA, Wavelength-Division Multiple Access
(WDMA) is another optical multiple-access scheme that has been attracting atten-
tion. As listed in Table 2.1, these three schemes have their pros and cons; which
scheme to use depends on the system environment [10, 33, 39].
As shown in Table 2.2, based on th e choices of signaling format, detection
method, synchronization requirement, coding domain, and code format, optical cod-
ing schemes can be generally classified into seven main categories:
1. Temporal amplitude coding (1-D)
2. Temporal phase coding (1-D)
3. Spectral amplitude coding (1-D)
4. Spectral phase coding (1-D)
5. Spatial-temporal amplitude coding (2-D)
6. Spectral-temporal amplitude coding (2-D)
7. Spatial/polarization-spectral-temporal coding (3-D)
51
52 Optical Coding Theory with Prime
TABLE 2.1
Comparisons of Common Optical Multiple-Access Schemes
Advantages Disadvantages
O-TDMA • Dedicated channels • Inefficient in bursty traffic
• Suitable for continuous traffic • System-wise synchronization
• High throughput • Hard limit on the number of
• Deterministic access active and possible users
WDMA • Dedicated channels • Inefficient in bursty traffic
• Suitable for continuous traffic • Need wavelength management
• High throughput and stabilization
• Deterministic access • Hard limit on the number of
• (Time) asynchronous access activ e and possible users
O-CDMA • Efficient in bursty traffic • Collisions
• (Time) asynchronous access • Random, statistical access
• Soft limit on the number of • Performance degrades with
active and possible users number of acti ve users
• Flexible user allocation
Source: Reproduced with permission from Guu-Chang Yang and Wing C. Kwong, Prime Codes with
Applications to CDMA Optical and Wireless Networks,Norwood,MA:ArtechHouse,Inc.,2002.
c
!
2002 by Artech House, Inc.
The first two categories involve one-dimensional ( 1- D) coding in the time domain.
The schemes in category 1 are based on incoherent signal processing and detection
that operate on the intensity of optical pulses; the optical pulses are spread out in
time [1–8, 33, 34, 40–42]. While the schemes in this category are the easiest of all
categories to implement an d (time) asynchron ous in gener al,theyrequiretheuse
of 1-D pseudo-orthogonal binary (0,1) codes, such as optical orthogonal codes and
the asynchronous prime codes in Chapter 3 [33, 43–47], with low cross-correlation
functions. On the other hand, by applying code synchronism, avariationintemporal
amplitude cod ing allows the use of shifted 1-D pseudo-or th ogonal codes, such as
the synchronous prime codes in Chapter 4, in order to improve cod e cardinality and
performance at the expense of system-wise synchronization [2, 33, 40]. Borrowing
aconceptfromwirelesscommunications,theschemesincategory 2 utilize optical
fields by applying 0 or
π
phase shifts to time-spreading optical pulses [10–13]. Using
coherent signal processing and detection, the schemes in this category allow the use
of orthogonal bipolar (−1, +1) codes, such as max imal-leng th sequences and Walsh
codes [48, 49], with zero (in-phase) cro ss-corr elation functions fo r minimizing mu-
tual interference. However, perfect code synchronization is required to maintain the
orthogonality of bipolar codes. In addition, phase coding requires special fibers to
preserve phase information during transmission. Nevertheless, these two 1-D tempo-
Optical Coding Schemes 53
TABLE 2.2
Classifications of Optical Coding Schemes
Signaling format Optical intensity Optical field
Detection method Incoherent Coherent
Synchronization Asynchronous Synchronous
Coding domain Temporal Spectral Spatial Polarization
Code format Unipolar (0,1) Bipolar (−1,+1)
ral coding techniques require a very long code leng th in ordertosupportasufficient
number of simultaneous users and possible subscribers. As a result, ultrashort optical
pulses and high-speed electronics are required. The schemesinbothcategoriesare
susceptible to fiber dispersion and nonlinearities.
In categories 3 and 4, 1-D amplitude and phase coding are performed in the
wavelength domain, respectively [15–21, 35, 50–52]. The spectral nature o f optical
codes is decoup led from the tempo r al nature of data; cod e length is now independent
of data rate, and system-wise (time) synchronization is not needed. A broadband
optical pulse is first dispersed into multiple wavelengths (or so-called wavelength
spreading), spectral coding is then performed by passing these wavelengths throug h
intensity or phase modulators, and the coded wavelengths arefinallyrecombined
to form a spectral codeword. Code length is determined by the resolution o f the
wavelength-spreading and coding devices, which, in turn, limit the number of pos-
sible subscribers because the cardinalities of the optical codes in use are usually
related to code length. With a special decoder design, orthogonal bipolar codes, such
as maximal-length sequences and Walsh cod es, can be used in both categories for
minimizing mutual interference. Furthermore, binary (0,1) codes, wh ich have a low,
fixed, in-phase cross-correlation function, such as the synchr onous prime codes in
Chapter 4, can also been used in spectral amplitude coding with a special decoder.
To support many users in temporal amplitude coding, 1-D binary (0,1) codes,
such as the optical orthogonal codes and prime codes (in Chapters 3 and 4), have
been designed with good correlation properties—thumbtack-shape autocorrelation
and low cross-correlation f u nctions—in order to optimize the discrimination be-
tween the correct-address codeword and interfering codewords. So, good 1-D bi-
nary (0,1) codes need to have long code length in order to spread out the effect
of mutual interference. In other words, very large bandwidthexpansionandhigh-
speed coding hardware are required to support sufficient numbers of simultaneous
users and possible subscribers. One possible way to lessen this problem is to uti-
lize 2-D coding, such as adding space or wavelength as the second coding dimen-
sion to the 1-D temporal-amplitude (or so-called time-spreading) codes. The spatial-
temporal amplitude coding schemes in category 5 allow the useoffreespace,mul-
tiple fibers, or multicore fibers with 2-D binary (0,1) codes being transmitted in the
time and space domains simultaneously [22–27, 53]. The spectral-temporal ampli-
54 Optical Coding Theory with Prime
tude (so-called wavelength-hopping time-spreading or, in short, wavelength-time)
coding schemes in category 6 require coding in the time an d wavelength domains,
in which fast wavelength hopping is involved in the pulses of 2-D binary (0, 1)
codes [14, 28–30,33, 34, 36,37, 54–61 ] . In general, 2 -D codes, such as the 2-D prime
codes in Chapters 5 and 6, provide lower probab ility of interception and offer coding
scalability and flexibility due to the use of two coding dimensions. These features
in the physical layer can be beneficial in supporting time-sensitive obscure transmis-
sions in strategic or military systems, where real-time encryption delay is critical and
software encryption at high speed is rather dif ficult [62].
To further reduce code length or improve code performance, the combination of
three coding dimensions has been proposed in category 7. For example, the three-
dimensional (3-D) spatial-spectral-temporal coding schemes in [32, 63] transport
wavelength-time pulses via multiple fibers with the use of 3-Dunipolarcodes,such
as the 3-D prime codes in Chapter 8. The 3-D polarization-spectral-temporal coding
scheme in [31] carries wavelength-time pulses along with thetwopolarizationsof
light via an optical fiber.
Future coding-based optical systems and networks are expected to support multi-
media services with different bit rates, qu ality-of-services (QoSs), and priority. For
instance, the use of specially designed 1-D and 2-D binary (0,1 ) cod es with multi-
ple lengths and variable weights but fixed low cross-correlation functions has been
proposed for supporting these types of multimedia services [33, 34, 64–68]. By using
the multilength prime codes in Chapter 7, one system clock andlaserswiththesame
pulse-width (or so-called chip-width) can be used for all services, simplifying system
hardware and timing requirements. Also, studies have shown that shorter codewords,
which are assigned to high e r bit-ra te services, have better code perfor mance and, in
turn, higher service priorities—an inherent characteristic in multilength coding.
To achieve high bit rate or spectral efficiency in coding-based optical systems
and networks, th e co ncept of m u ltiple-bit-per-symbol transmission has been intro-
duced by means of pulse-position modulation, multicode keying, and shifted-code
keying [69–76]. The advantages of symbol transmission are threefold: the effective
bit rate is increased by the number of bits per symbol, in essence, trading hard-
ware complexity for reducing bandwidth expansion and electronic speed; spectral
efficiency is improved; and user code obscurity is enhan ced because bit 0s ar e also
transmitted in codewords and eavesdroppers cannot determine the transmission of
bit 0s or 1s by simply detecting the absence or presence of optical intensity in the
downlink fiber [62]. The prime codes that are suitable for multicode keying and
shifted-code keying are also studied in this book.
The rest of this chapter is organized as fo llows. The coding techniqu es and en -
abling hardware technologies of the seven categories of optical coding schemes are
reviewed in Sections 2.1 through 2.7. The special technique of supporting multi-
rate, multimedia services in coding-based optical systems an d networks by means
of multiple-length codes is studied in Section 2.8. Afterward, Section 2.9 introduces
multicode keying and shifted-code keying for increasing bitrate.Additionaldesigns
of coding devices, based on arrayed waveguide gratings [57–60,77] and fiber Bragg
Optical Coding Schemes 55
gratings [78–82], are investigated in Section 2.10. Finally, various potential applica-
tions of optical coding, in addition to O-CDMA, are discussedinSection2.11.
2.1 1-D TEMPORAL AMPLITUDE CODING
transmitter
receiver
recovered
electrical data
one station
threshold
detector
photo-
detector
optical
decoder
electrical data bits
intensity
modulator
optical
encoder
electrical path
optical path
:
:
:
:
:
:
:
:
star
coupler
optical
pulses
FIGURE 2.1 Atypicalcoding-basedopticalsystemmodel.(Source:Reproduced with per-
mission from Guu-Chang Yang and Wing C. Kwong, Prime Codes with Applications to CDMA
Optical and Wireless Networks,Norwood,MA:ArtechHouse,Inc.,2002.
c
! 2002 by Artech
House, I nc.)
Figure 2.1 outlines the basic configuration of a typical coding-based optical system
and network. It consists of multiple stations (or users) linked to a shared optical
medium via optical fibers or free space [1, 33]. The shared medium usually consists
of an optical multiplexing/demultiplexing device, such as astarcoupler.Thedevice
is used to combine optical codewords from and then distributetoallstations.Each
station consists of a pair of optical transmitter and receiver, whose structur es depend
on the coding scheme in use.
In incoherent on-off keying (OOK) modulation, a u ser sends out an optical code-
word corresponding to the address (signature) codeword of its intended receiver for
adatabit1,butnothingforadatabit0.Figure2.2(a)showsthe timing diagrams of
acontinuousstreamofopticalclockpulsesofwidthT
c
and repetition rate 1 /T.Fig-
ure 2.2(b) shows an example of low-bandwidth nonreturn-to-zero electrical data bits
of period T at one station in Figure 2.1 . Electro-o ptic OOK conversion isperformed
using the voltages of data bits to control the opening and closing of the intensity
modulator in the transmitter [1, 3 3]. Every data bit 1, which carries a hig h voltage,
will close the switch and let pass one clock pulse, resulting in a high-bandwidth data-
modulated optical signal, as shown in Figure 2.2(c). Assume that 1-D time-spreading
binary (0, 1) codewords ar e used as the add r ess codewords of the stations. To accom-
modate the 1s and 0s of the codewords, each bit period T is subdivided into a number
of time slots (or so-called chips) of width T
c
,givingcodelengthN = T /T
c
.Following
the data-bit pattern in this illustration, Figure 2.2(d) shows the corresponding time-
spreading binary (0, 1) codewords generated by an optical en co der. Finally, these
codewords are multiplexed with the codewords from all stations at the star coupler,
as exemplified in Figure 2.2(e).
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.