2010 IEEE 24th International Conference on Advanced Information Networking and Applications Workshops

Adaptive Channel Coding and Modulation

Scheme Selection for Achieving High
Throughput in Wireless Networks
1 3
Atif Sharif , 2Vidyasagar M. Potdar Rana Fayyaz Ahmad
Digital Ecosystems and Business Intelligence Institute Digital Ecosystems and Business Intelligence Institute
Curtin University of Technology Curtin University of Technology
Perth, Western Australia Perth, Western Australia
1 3
[email protected], [email protected]
2
[email protected]

Abstract—Modern wireless communication demands reliable Channel coding is one of the main tools that increase the
data communication at high throughput in severe channel transmission reliability at higher data rates. The purpose of the
conditions like narrowband interference, frequency selective channel coding scheme is to guard the information against
fading due to multipath and attenuation of high frequencies. channel disturbances, achieved by introducing data redundancy
Traditional single carrier systems address this set of problems by for error detection and correction. In wireless communication
the use of complex, computationally intensive equalization filters. systems erroneous signal detection and its correction can be
The Orthogonal Frequency Division Multiplexing (OFDM) based achieved by the use of proper error control coding techniques.
system, as opposed to single-carrier systems, is considered to be The bursty traffic sources which are multiplexed statistically,
the future of the wireless communication and is being used to
the fast and dynamic channel capacity allocation can be
achieve high data rate by overcoming severe channel conditions
without the use of these complex filters. This paper discusses the
achieved by Single-carrier Quadrature Phase Shift keying
problem of Adaptive Modulation scheme selection through an (QPSK) or Quadrature Amplitude Modulation (QAM) [18].
OFDM based system over parallel frequency selective fading However such single-carrier systems show their effectiveness
channels. An adaptive coding scheme is proposed by using in severe multipath applications requiring intensive processing
Generalized Concatenated Codes (GCC), which have simple at high data rates [18] only if they use adaptive equalizer.
structure and are designed in such a way that they are best suited Another problem with single carrier systems operating is
for fading channels. GCC are based on binary cyclic codes. The susceptibility to interference.
criterion of the proposed research is to optimize the throughput
of a wireless system. Depending on the quality of sub-channels an A. Introduction to OFDM
adaptive modulation selection scheme and code assigning method
Orthogonal frequency division multiplexing (OFDM)
is proposed. The proposed research combats against channel
impairments better than those used in conventional systems by
sometimes called multi-carrier or discrete multi-tone
exploiting individual sub-channel condition. Results show better modulation used for digital TV in Japan, Europe and Australia.
performance in terms of higher throughput by minimizing the bit OFDM, being spread spectrum by nature, uses a number of
error rate. carriers, which are spectrally distributed at precise frequencies,
for data transmission. Orthogonality is being provided by this
Keywords- GCC, Channel State Information (CSI), OFDM, spectrum spacing which prevents demodulator interference
COFDM, Fading. with other frequencies. OFDM offers a number of benefits like
resiliency to Radio Frequency (RF) interference, high spectral
I. INTRODUCTION efficiency, and lower multi-path distortion. The increasingly
need for high speed mobile data transmission is a byproduct of
Wireless communication technology, a fundamental part of the recent advancements in the digital wireless communication.
modern information infrastructure, is growing day by day. To counter this increasingly need for high speed data
Spectrally efficient and reliable transmission schemes are used transmission capacity researchers are continually engaged in
for high speed transmission data rate over wireless link. envisaging new modulation techniques. As power is the biggest
However reliable data communication is one of the major concern in most of the wireless communication, OFDM has
problems in modern wireless channels, since these channels feasibly gained its high processing power at economical cost in
should be able to tolerate the effects of signal fading channels comparison to existing wireless technologies. And now OFDM
and Inter Symbol Interference (ISI). Mobile data is considered as major technological contender for 4G wireless
communication is even more challenging. It requires applications. DSL, HDTV broadcasting, DAB, IEEE802.11 are
continuous measurement and updating the Channel State some of the key examples of applications involving OFDM for
Information (CSI) and adaptation of the coding/modulation high rate wireless communication. In OFDM high bandwidth
techniques according to the new environmental conditions. It efficiency is achieved by using the orthogonal tomes for signal
also involves proper power distribution techniques [1]. modulation. The tones are evenly spaced at defined intervals,

978-0-7695-4019-1/10 $26.00 © 2010 IEEE 200

DOI 10.1109/WAINA.2010.197
equal to symbol rate, in frequency band and can be easily RA =kA/ n (2.1)
separated at the receiving end. In case of mobile wireless RB =kB/m (2.2)
channels multi path delay is a serious issue and is being
handled effectively by OFDM which simultaneously handles The product code C is obtained from the codes A' and B' in
the symbol duration and data transmission in parallel. The the following manner:
channel capacity in case of OFDM based system is increased 1. Place kA x kB information bits in a metrics having kA
due to the fact that OFDM splits the data bearing the radio columns and kB rows.
signal into multiple data blocks. Each data block in turn being
modulated by the different sub carriers which are orthogonally 2. Code A' will be used to code kB rows resulting in an
spaced in frequency spectrum thereby preventing the array having n columns and kB rows.
interference occurred during the simultaneous transmission 3. Code B' will then be used for coding the n columns.
over wireless channel. In OFDM the spacing between the sub
carriers is chosen in such a way that they are all orthogonal to The construction of the product code C with constituent
each other. In [14] researchers envisaged a method for codes A' and B' is illustrated in Figure 1 [1].
maintaining this orthogonality. According to the method the The parameters of the resulting product code will be [mn,
sub carrier’s frequency spacing is chosen to the inverse of the kAkB, dAdB] and its rate will be equal to RARB. Therefore,
sub carrier symbol duration. we can construct long block codes starting by combining two
The rest of the paper is organized as follows. Section II lists short codes. Product codes have much smaller minimum
some of the related work followed by section III where we distance as compared to optimal codes. The error correcting
have outlined the problem definition. The Design parameters potential of product codes is quite large. They have burst error
and the conceptual framework are discussed in the Section IV correction capability. The covering radius of product codes is,
and V respectively. Section VI shows the effectiveness of the usually, much greater than half the minimum distance of the
proposed scheme followed by the Section VII where we have code [10]. Forney was the first to propose concatenated codes
discussed the future research directions and finally the [4]. The aim was to find a code that approaches the channel
conclusion of the paper is drawn in Section VIII. capacity with practical decoding complexity that increase
polynomially with the length of the code [11]. The proposed
code was a concatenation of a relatively short inner code with
II. RELATED WORK
an outer long Reed-Solomon code [12].A very important
Concatenated codes are used to obtain long and powerful characteristic of concatenated codes is the Hamming weight
codes by using simple constituent codes. Product codes are a distribution [19] of the codes and their Euclidean distance
kind of concatenated codes and are presented by Elias [3]. spectrum [17]. These characteristics are very important for
They can be represented as a set of matrices such that each row determining the error correction capability of the codes by
in these matrices is a codeword in one constituent code and using different bounds. Generalized concatenated codes are
each column is a codeword in another constituent code. One further generalization of concatenated codes. They can be
important property of product codes is burst error correction. viewed as binary matrices with the rows and columns
belonging to many different block codes. They were introduced
by Zinoviev [5]. Concatenated codes are efficient in wireless
communication channels because they have comparatively high
minimum distances and they have proper structure for burst
error correction without the need for extra interleaving.
Interleaving is used to transform burst errors into random errors
which then can be corrected by forward error control codes.
Concatenated codes have higher reliability with slight increase
in complexity. By using concatenated codes different data rates
can be achieved as several different codes are used for
encoding the rows and columns instead of restricting oneself to
only one code for the rows and one code for the columns.
Adaptive modulation has also recently gained momentum in
wireless systems because it provides more efficient use of
channel. For OFDM systems the researchers in [7] and [13]
investigated the adaptive modulation criteria. The proposed
Figure 1. Structure of Product Codes [1] research in [9] had taken the time varying channel conditions
for adaptive modulation and coding. For OFDM systems, the
Product codes are serially concatenated codes [12]. The adaptive modulation is investigated according to Average
idea of product codes is quite simple and powerful at the same Channel Conditions by [8]. The idea of application in OFDM
time where very long block codes can be constructed by the use system is presented in [2]. For OFDM systems, researcher in
of two or more shorter constituent codes. Consider A' and B' as [1] had proposed an adaptive coding scheme that uses
two block codes with performance parameters [n, kA, dA] an generalized concatenated codes. The proposed coding scheme
[m, kB, dB], respectively [1]. The rates of the codes A' and B' was applied to HIPERLAN/2, whose performance was then
are denoted by RA and RB, respectively, and are equal to: analyzed against the standard convolution coding scheme in

201
use. The objective of research was to enhance the system’s in mutually deciding the channel modulation scheme for
throughput and was achieved by adaptively selecting the gaining increased performance which results in high data rate
modulation and channel coding schemes based on varying sub- as compared to non adaptive coded schemes.
channels conditions.
Rate Adaptive Transmission design feature relies more on
the quality of the wireless channel and it ensures the maximal
III. PROBLEM DEFINITION use of the transmission bandwidth. It achieves high spectral
The modern wireless communication demands reliable data efficiency by increasing the sending information rate during the
communication at high throughput in severe channel conditions good channel conditions and vice versa. Channel status in form
like narrowband interference, frequency-selective fading due to of channel state information (CSI) is being sent by the receiver
multipath and attenuation of high frequencies. The existing to the transmitter on the reverse wireless link. This information
Single carrier systems address the above mentioned problems will then be used by the transmitter in deciding the rate plan for
by involving complex and computationally intensive channel the next transmission. This rate adaptive transmission ensures
equalizers. Also reliable data transmission and security is of the maximal utilization of the channel bandwidth for an
major concern these days. Thus a need to make more efficient acceptable Bit-Error-Rate (BER).
use of bandwidth transmissions without creating situations The Severity of Wireless Communications is much greater
where signals would be subject to above mentioned than that for AWGN channels. For AWGN channels, good
phenomenon is generally desired. In order to achieve these codes can be designed by making the distance between the
design objectives of wireless data communication there are code words as large as possible, or, alternatively, to make the
some key design considerations like channel coding, Multiple number of code words at small Euclidean distance as small as
access schemes, adaptive modulation, rate adaptive possible. But, for fading channels, the severity of channels is
transmission etc which must be taken into considerations and mainly tackled by the diversity in communication. Diversity in
are discussed in the following section. communication means sending a copy of the message, or part
of a copy, on different paths or channels in order to increase the
IV. DESIGN CONSIDERATIONS reliability. It can be achieved by diversity in carrier frequency,
Following parameters must be considered while designing space diversity, by coding or a combination of all methods.
the wireless data communication system, these are
V. PROPOSED GENERALIZED CONCATENATED
1. Channel Coding
CODE SCHEME
2. Multiple access Schemes In this paper, the problem of Adaptive Coded Modulation
3. Adaptive Modulation through an OFDM based system over parallel frequency-
selective fading channels using Generalized Concatenated
4. Rate-adaptive Transmission Codes (GCC) is proposed. These codes are simple by structure
5. Severity of Wireless Communication and rely on the concatenation of two or more shorter length
codes. The large diversity design feature of such codes makes
Channel coding is also considered as an essential and them choice of interest for use in case of channels suffers from
important element of communication system, one of the main fading also called fading channels. GCC are based on binary
tools that increase the transmission reliability at higher data cyclic codes. The decoding of the proposed codes is performed
rates. The purpose of the channel coding scheme is to guard the by GMD decoding. These codes can be used in current wireless
information against channel disturbances during transmission, systems where approaching channel capacity is, usually, of less
and this can be achieved by introducing data redundancy for importance than the processing delay, energy consumption or
error detection and correction. It protects data against integrated circuit chip size [1]. The method used for designing
transmission errors to ensure adequate transmission quality (bit the codes bear great similarities to the ideas presented by in [1]
or frame error rate). It is power efficient as compared to the with some modifications that take into consideration the
uncoded case, the same error-rates are achieved with much less channel conditions.
transmit power at the expense of a bandwidth expansion. Error
control coding has gained fundamental importance in wireless A. Proposed Coded OFDM (COFDM)
communication systems and deals with techniques for detecting
and correcting signal errors. In this Paper, COFDM is used. First the channel is
subdivided into N number of sub-channels. Depending upon
Multiple access schemes narrow or wide band, are a very the channel state information of the sub-channels the carrier to
important means of allowing a large number of mobile users to interference ratio and the AWGN noise is calculated. Based on
access a finite amount of frequency spectrum simultaneously. this information the modulation and the encoding are
By doing this capacity can be increased. In simple words they preformed. Thus the COFDM exploits the channel state
ensure the efficient use of the frequency spectrum. They are information in order to achieve the best results that would
also preferred over Radio spectrums since they are very ensure the highest throughput. Here, the concept of multiple
expensive. accesses in OFDM is exploited to accommodate multiple codes
with data rates that vary in coding and in modulation. Thus it
Adaptive modulation schemes take channel conditions into
would be adaptive multiple access on the OFDM. The
account at both the transmitter and receiver end. And this
following points taken in consideration
knowledge of channel conditions helps transmitter and receiver

202
 1. A coding scheme, generalized concatenated codes, with The above two aspects are interrelated in the sense that the
good error correction capabilities especially in fading channels. choice of one effects the others and vice versa. The criterion is
to maximize system’s throughput by maintaining probability of
2. Choosing optimum Coding dimensions according to the error.
individual channel conditions and Proper selection of codes for
row blocks. 1) Modulation Scheme
3. Designing an effective adaptive coded modulation The choice of modulation scheme plays a very important
algorithm for each sub-channel such that throughput is role. This is so because it would be a balancing act between the
maximized, while keeping the complexity of the system to a throughput and the probability of error. This means that the
minimum. order of modulation needs to be selected in such a manner that
the maximum allowable probability of error is never exceeded.
4. Generalized Minimum distance (GMD) decoder will be However, the catch is the throughput must be maximized.
used for decoding. In the proposed scheme the criterion of sub- When this is done, there would be a compromise with the
channels quality is being used in defining the adaptive probability of error. Depending upon the (Channel State
modulation scheme selection and the corresponding Information) CSI the best possible modulation scheme is
information coding scheme that adapts according to the varying selected. The following discussion will focus upon the fact that
wireless channel conditions. The proposed research offers the CSI has been calculated; and the proper modulation order
significant advantage in protecting the information as and type need to be predicted. The modulation schemes that
compared to conventional schemes that did not take into are used here are discussed in this section. Each sub-channel is
account the conditions of individual wireless channel. The Rayleigh in nature. So the processing of sending data on any
criterion of study is the optimization of the throughput of the one of the channel can be represented by the multiplication of
system, while maintaining the PER (Packet Error Rate). the data and the channel state information and then adding the
AWGN noise. This process can be represented by the
B. System Overview following equation
Instead of the average standardized one, the structure of the r(t) = c(t).s(t)  n(t) (5.1)
proposed GCC code has taken into account the dependency of
individual sub-channels’ condition. It is shown in Figure.2 [1]. Where, r(t) is the received signal, c(t) is the Channel State
Information, s(t) is the data signal and is the AWGN noise.
This Channel State Information (CSI) and the AWGN noise is
used to determine the modulation order. For this, the water-
filling algorithm is used to calculate the Carrier-to-Interference
ratio. This Carrier-to-Interference ration will used to calculate
the theoretical probability of error. There are four types of
modulations used here. These methods are
x BPSK
x QPSK
x 8QAM
x 16QAM
Figure 2. System Structure [1]
x 64QAM
C. Channel Estimation:
In order to select the best possible parameters we need to
know the Channel State Information. This CSI will determine
the criteria that would allow us to select the right parameters
for transmitting the data. This CSI has to be reliable in order to
produce good results. In wireless communication systems, the
channel estimation is performed by transmitting a training
sequence [20] [21], i.e., a sequence of symbols that are known
to the receiver in advance. This sequence of symbols is used to
estimate the parameters of the channel. There are basically two
aspects which are taken into account while system structuring
these includes
A) Choosing different modulation scheme for each sub
channel
B) For each individual sub-channels a decision on selecting Figure 3. Bit Error Probability of Channel
the codes blocks

203
 Using the above equation (2.2) their behavior can be where the practical probability of error is Pb.
depicted in the Figure 3. From this graph we need to choose a
point that will serve as threshold point for the maximum 2) Coding Scheme
allowable Bit Error Rate. From the above figure, we can see Channel coding is used in communication systems to
that QPSK is most robust of them all. Therefore, by choosing a eliminate or greatly reduce the errors that are introduced by the
maximum probability of error for this method will provide us channel, so that a reliable communication can be possible. All
with a threshold for a minimum operate able SNR. In our case codes are, basically, a pre-selected subset of sequences from
let the maximal allowable rate is 10-3. the total space of signal sequences. A simple GCC type coding
scheme is implemented where the row vector will be loaded by
This performance would be possible only for the SNR the different information bits according to the code used. After
greater then 10dB at least. From the graph it can be seen that this different binary Bose-Chaudhuri-Hochquenghem (BCH)
there is no modulation method that would be able to perform codes will be applied to achieve the row encoding. Each row
better than the QPSK for a given probability of error. Therefore code is subset of each other. After the row encoding the one
we will assign this method whenever the channel is very bad binary BCH code will be used to perform the column encoding.
and there is no possible way for the increase of throughput Figure 3 depicts the structure of the GCC used in this system,
without compromising the system performance with respect where the colored and shaded area points to the information
error probability. The choice of these modulation methods and redundancy bits. The information required in order to build
depends on two reasons the Generalized Concatenated Code is the modulation scheme.
1. Their ability to provide higher throughput with Since in our system OFDM is used there will be N number of
acceptable bit error rate sub-channels. Each of these sub-channels will have its own
modulation scheme. The algorithm for the adaptive selection of
2. Their contribution to the construction of GCC’s the modulation schemes adaptively computes this modulation
Each sub-channel will be assigned one of these modulation scheme. Once it has done, we have the following information
schemes. This process depends a lot on each sub channel’s x Method of Modulation
Channel State Information (CSI). The Water-Filling algorithm
to compute the Carrier-to-Interference ratio (C/I) will use the x Order of Modulation
CSI. Thus (C/I) of each sub-channel will provide the criteria
for the selection for the method that would provide us with the x Numbers of sub channels
best throughput. The carrier to interference ratio will provide us 3) Algorithm
an upper limit on the probability of error for each method. For The following algorithm will demonstrate how the
this the following formulae are used: modulation scheme is selected for each of the sub-channels
along with the respective rates. The process of power allocation
is also covered within these steps. The proposed concept is
§ C· (5.2)
Pbth Q ¨¨ 2 * ¸¸ forBPSK summarized below.
© I ¹
a) STEP 1: POWER DISTRIBUTION AMONG
§ C· SUBCHANNELS:
Pbth Q ¨¨ ¸ forQPSK
¸
(5.3)
© I ¹ 1. Compute the SNR of each sub-channel as Carrier-to-
interference ratio C/I
§ 3 ·§¨ §¨ C ··
Pbth ¨ ¸¨ Q ¨ 2 * ¸¸ ¸¸ for16 QAM (5.4)
2. Compute the initial power distribution for all the sub-
© ¹© ©
4 I ¹¹
channels
§ 7 ·§¨ §¨ C ··
¸ ¸ for 64 QAM (5.5) 3. Check if any of the sub-channels is allotted negative
Pbth ¨ ¸¨ Q ¨ 2 * ¸¸
© 12 ¹© © I ¹¹ power. If not then follow to the next step otherwise follow the
following procedure:
These formulae will give us the theoretical bound on the
a) Those indexes that have a positive value of power will be
probability of error. This bound is computed for each sub-
marked.
channel for each of these methods. Once this is computed then
a comparison is preformed in order to determine which b) These marked indexes will refer to those values of C/I
modulation scheme to be used. For this purpose the practical that correspond to positive values of sub-channel power. Thus
probability of error is calculate for every sub-channel using power will be redistributed among these marked C/I sub-
each of the above mentioned modulation schemes. After this is channels while ignoring those with negative power.
done, a comparison is preformed. In this comparison, the
method that would yield the maximum throughput is selected. c) This negative power test is repeatedly preformed until no
In the worst case scenario QPSK would be selected so as the negative power is assigned to any channel.
probability of error never exceeds the allowable level. So in d) If there is no more negative sub-channel power then go
order to select the right modulation scheme the following to the step (following) 4.
criteria
 Pb d Pbth (5.6)

204
 e) The computed values of C/I and initial values of
allocated power per sub-channel are set as final values for
further computation.
b) STEP 2: COMPUTE Pb FOR EACH SUB-CHANNEL:
Using the computed C/I of each sub-channel the empirical
values of the probability of error is computed. Along with this
the practical probability of error is also computed for each sub-
channel. The above steps are preformed for each of the four
modulation schemes.

Figure 5. Comparison of Throughput

Figure 6. Comparison of BER

In this paper after a large number of iteration the set of

horizontal encoders that was selected are kA =
Figure 4. Algorithm {30,36,39,45,51,57} and for the vertical encoder kB =57 and n
= m = 63 . Here, we seeked no optimum solution by the use of
c) STEP 3: CODING AND MODULATION SCHEME sub-channels for redundancy. Figure 5 depicts the throughput
ASSIGNMENT PER SUB- CHANNEL: comparison of the proposed idea and standardized H/2 one
Compare the empirical and practical values of error [15]. It also figures out the case of six sub-channels for
probability for each sub-channel selecting the best scheme for redundancy [1]. The BER is depicted in Figure. 6, here
maximization of throughput Based on the modulation scheme discontinuities are a result of the change of the modulation
select respective coding scheme. scheme used.
d) STEP 4: LOAD THE INFORMATION BITS AND It is worth mentioning that high bandwidth, due to frequent
MODULATE. reporting processes, for reverse link will be required in the
proposed improvement scheme. Additionally the process
VI. RESULTS complexity measured in terms of number of cyclic operations
per decoding, is nearly comparable to the conventional
The main idea while erecting the simulation, that uses decoding process used in commercial OFDM systems [16].
Matlab as simulation environment, to support the acclamations,
the main idea was to keep the whole process as simple as
A. Comparisons between Referenced and Proposed
possible. At the same time, it was also essential to imitate the
reality as much as possible. This purpose involved some Predicament
assumptions. It was understood that there was a perfect The Figures 5 and 6 illustrates that the referenced adaption
knowledge of sub-channels conditions. of H/2 shows better performance in terms of throughput, while

205
sustaining a low PER. It is also possible to attain a certain VIII. CONCLUSION
throughput at low power there by resulting in less interference Proposed algorithm result in minimum system complexity
and hence increased system’s capacity. Also the proposed and results in great performance gains as compared to
enhancement outperforms H/2 in certain applications which are conventional systems. Here, during the deep fading of the
subject to tight error constraints (< 1%). The decoding process channel, transmission is closed. So it is possible to obtain some
complexity is measured in terms of the cycles/operations it gain in total rate by switching off the transmission when the
would take to decode, is much less or at worst as compared to quality of the channel degrades below a certain level. This is a
the conventional decoding process used in commercial OFDM simple form of adaptation similar to that presented by
systems. Here, the throughput is maximized while attaining the Goldsmith et al in [6]. However, it is maintained that the net
minimum probability of error. Therefore, as in general the gain in rate is very small especially for high signal to noise
proposition was able to obtain an optimally maximized ratios which is the same conclusion in [6].GMD decoding of
throughput. generalized concatenated codes requires very low complexity.
This decoding method decodes all errors of weight less than
VII. DISCUSSION/ FUTURE WORK half the minimum distance. GMD decoding performs very well
In this paper, we have used five types of modulations if the minimum distance of the concatenated code is very large,
namely BPSK, QPSK, 8QAM, 16QAM and 64QAM. These i.e., close to the maximum possible value. However, if the
five modulation schemes are used for the best utilization of minimum distance is not very large, most of the error
channels. In reference paper [1] number of rows in GCC was correction potential lies in correcting error patterns slightly
equivalent to number of channels, so there was not enough greater than half the minimum distance [1]. Proposed algorithm
redundancy added. In proposed work, numbers of row codes result in minimum system complexity and results in great
are equivalent to the number of channels, so enough performance gains as compared to conventional systems.
redundancy is added to protect data from any corruption. So
here data is better protected as compared to the conventional REFERENCES
systems. In general we have used the Adaptive Modulation [1] Omar-Al-Askary et al, “Coding and iterative Decoding of Concatenated
scheme selection in this paper in order to use each channel Multi-level Codes for Rayleigh Fading Channel,” Doctoral Project in
condition to its best, while maintaining the acceptable level of Radio Communication Systems, S3, KTH, Stockholm, Sweden, 2006
probability of error and maximizing the throughput. Thus, [2] L. Hanzo et al, “Adaptive Multicarrier Modulation: A Convenient
resources are not wasted and the COFDM is more efficiently Framework for Time-Frequency Processing in Wireless
Communications,” IEEE Proceedings of the IEEE 88, May, 2000
used. This is because the adaptive modulation not only
increases the throughput but also helps to encode the OFDM to [3] P. Elias. “Error-free coding,” IEEE Transactions on Information Theory,
4:29– 37, 1954.
better suit the channel conditions i.e. the Channel State
[4] David G. Forney Jr. “Concatenated Codes,” The M.I.T press, Cambridge
Information, Carrier-to-Interference ratio and the SNR of each Massachusetts, 1966.
sub-channel. From the figures shown in result section, the [5] V. A. Zinov’ev. “Generalized cascade codes,” 1976.
proposed predicament that the results of the modification are
[6] A. Goldsmith and P. Varaiya. “Capacity of fading channels with channel
better in both aspects that is the throughput and the probability side information”. IEEE Transactions on Information Theory, 43:1986 –
of error as well. Thus the modification better utilizes the 1992, Nov. 1997.
channel condition. We have been able to achieve these results [7] O. A. Al-Askary et al, “Subcarrier based adaptive modulation in
because in state of deep fading the transmission would hiperlan/2 system,” In 2004 IEEE International Conference on
shutdown. Since a main focus was to improve our throughput; Communications, pages 3460–3464. IEEE, IEEE, Jun. 2004.
it was observed that whenever the channel was in good [8] Omar Al-Askary et al , “Adaptive Coding for OFDM Based Systems
condition the modulation scheme with the highest rate would using Generalized Concatenated Codes,“ Radio Communication
Systems, S3, KTH, Stockholm, Sweden
be selected while maintaining the acceptable probability of
[9] H. Matsuoka et al, “Adaptive modulation system with variable coding
error. This in turn will adaptively select a coding scheme for rate concatenated code for high quality multimedia communication
each sub-channel. Here, the process complexity measured in systems,” Vehicular Technology Conference, Mobile Technology for the
terms of number of cyclic operations per decoding, is much Human Race, IEEE 46th, pages 487–491. IEEE, IEEE, Apr. 1996.
less or at worst as compared to the conventional decoding [10] G. Cohen et al “Covering Codes,” North Holland Math. Library, 1997.
process used in commercial OFDM systems [15]. Here number [11] David G. Forney Jr. “Performance and complexity,” In 1995 IEEE
of row codes are equivalent to number of channels, therefore International Symposium on Information Theory, page 1, IEEE, IEEE,
more redundancy is added. So the data is better protected ep. 1995.
against errors as compared to the conventional systems. Here [12] Florence Jessie MacWilliams et al, “The Theory of Error-Correcting
the arrangement of the sub-channels in a particular order is not Codes,” North-Holland, 1977.
required. Also there is no need to have a particular code [13] L. Zhen et al ” Link adaptation of wideband OFDM systems in multi-
path fading channel,” In proc. of IEEE Canadian Conference on
arrangement. Here adaptive coding and adaptive selection of Electrical and Computer Engineering, CCECE 2002, IEEE, May 2002,
modulation scheme according to individual channel conditions pp.1295– 1299.
results in better performance of GCC. GCC is build up on [14] Teng Li and M. Collins. “Capacity and Coding of Flat Fading Channels
linear binary codes and provide good performance under burst without Channel State Information”Department of Electrical
error conditions by eliminating the need of an Engineering University of Notre dame. Notre Dame, IN, 46556
interleaver/deinterleaver. These codes also proved to perform [15] Johan Malmgren. “Product code implementation in an EDGE like
well in fading channels. system: Decoding and performance”, M.sc., project, KTH, radio
systems, May 2003.

206
[16] P.Bansal and A.Brzezinski. “Adaptive Loading inMIMO/OFDM
Systems” December 13, 2001
[17] MADDOCKS, M.C.D. “An introduction to digital modulation and
OFDM techniques”, BBC Research Department Report No. RD
1993/10, 1993.
[18] M. Sellars and D. Kostas, “Comparison of QPSK/QAM OFDM and
Spread Spectrum for the 2-11 GHz PMP BWAS”, In response to the
Call For Contributions, by IEEE 802.16.3-00/07r1, 2000.
[19] J. D. Parsons. “The Mobile Radio Propagation Channel”. JohnWiley and
Sons, second edition, 2000.
[20] G. Forney. Us- patent #1971000179653: “Training adaptive linear
filters”, Mar.1973.
[21] S. Qureshi. “Fast start-up equalization with periodic training sequences”.
IEEE Transactions on Information Theory, 23(5):553 – 563, Sep. 1977.

207