|
DSpace@nitr >
National Institue of Technology- Rourkela >
Thesis (Doctor of Philosophy) >
Please use this identifier to cite or link to this item:
http://hdl.handle.net/2080/661
|
Full metadata record
| DC Field | Value | Language |
|---|
| contributor.author | Mohanty, Debiddutta | - |
| contributor.author | Panda, G (Guide) | - |
| date.accessioned | 2008-04-24T04:14:45Z | - |
| date.available | 2008-04-24T04:14:45Z | - |
| date.issued | 2008 | - |
| identifier.citation | Channel Equalization using GA Family, Doctoral Thesis submitted to National Institute of Technolgy Rourkela | en |
| identifier.uri | http://hdl.handle.net/2080/661 | - |
| description | Copyright for the thesis belongs to National Institute of Technology, Rourkela | en |
| description.abstract | High speed data transmissions over communication channels distort the trans-
mitted signals in both amplitude and phase due to presence of Inter Symbol Inter-
ference (ISI). Other impairments like thermal noise, impulse noise and cross talk
also cause further distortions to the received symbols. Adaptive equalization of
the digital channels at the receiver removes/reduces the e®ects of such ISIs and
attempts to recover the transmitted symbols. Basically an equalizer is an inverse
¯lter which is placed at the front end of the receiver. Its transfer function is inverse
to the transfer function of the associated channel.
The Least-Mean-Square (LMS), Recursive-Least-Square (RLS) and Multilayer
perceptron (MLP) based adaptive equalizers aim to minimize the ISI present in
the digital communication channel. These are gradient based learning algorithms
and therefore there is possibility that during training of the equalizers, its weights
do not reach to their optimum values due to the mean square error (MSE) being
trapped to local minimum. In other words true Weiner solution is not achieved
because of gradient based training. The bit-error-rate (BER) performance of the
equalizer further degrades when data transmission takes place through nonlinear
channels.
The standard derivative based algorithms su®er from local minima problem
while obtaining the solution of the weights. To prevent the premature settling of
the weights, evolutionary computing based update algorithm is proposed which is
essentially a derivative free technique. Equalization is basically an iterative process
of minimization of mean square error. Thus equalization can be viewed as opti-
mization problem. The minimization of squared error is achieved iteratively using
GA. Thus GA based approach is an e±cient method to achieve adaptive channel
equalization. In the present thesis classes of new adaptive channel equalizers are
proposed using derivative free evolutionary computing tools such as Genetic Algo-
rithm (GA) and Particle swarm optimization (PSO). These algorithms are suitably
used to update the weights of the proposed equalizers. The performance of these
equalizers is evaluated in terms of speed of convergence, computational time and
bit-error-rate (BER) and is compared with its LMS based counter part. It is ob-
served that the new set of adaptive equalizers o®er improved performance so far
as the accuracy of reception is concerned. However, in order of increasing training
time the equalizers may be arranged as the adaptive Genetic Algorithm (AGA),
Particle Swarm Optimization (PSO), Real coded Genetic Algorithm (RCGA), Bi-
nary coded Genetic Algorithm (BGA) based equalizer.
However being a population based algorithm, standard Genetic Algorithm
(SGA) su®ers from slower convergence rate. To minimize the training time three
di®erent adaptive GAs (AGAs) are proposed in the thesis and their convergence
times have been compared. The thesis also investigates on the new equalizers us-
ing Real coded Genetic Algorithm (RCGA) and Binary coded Genetic Algorithm
(BGA). Their performances are also evaluated.
In the conventional FLANN (Functional Link Arti¯cial Neural Network) [1]
structure the complexity increases due to incorporation of more number of paths
after functional expansions. To reduce the structural complexity some pruning
of structure is essential. Keeping this in mind the GA based pruning strategy is
used in the FLANN identi¯er. It is observed that about 50% of the total signal
paths can be pruned keeping the performance identical to that of original FLANN
structure. | en |
| format.extent | 5169924 bytes | - |
| format.mimetype | application/pdf | - |
| language.iso | en | - |
| publisher | NIT Rourkela | en |
| title | Channel Equalization using GA Family | en |
| type | Thesis | en |
| Appears in Collections: | Thesis (Doctor of Philosophy)
|
Files in This Item:
| File |
Description |
Size | Format |
|---|
| thesis-final-dmohanty.pdf | | 5048Kb | Adobe PDF | View/Open |
|
Show simple item record
All items in DSpace are protected by copyright, with all rights reserved.
|
