Content uploaded by John Sum
Author content
All content in this area was uploaded by John Sum on Apr 07, 2015
Content may be subject to copyright.
A preview of the PDF is not available
Improved transmission of vector quantized data over noisy channels
Abstract and Figures
The conventional channel-optimized vector quantization (COVQ) is very powerful in the protection of vector quantization (VQ)
data over noisy channels. However, it suffers from the time consuming training process. A soft decoding self-organizing map
(SOM) approach for VQ over noisy channels is presented. Compared with the COVQ approach, it does not require a long training
time. For AWGN and fading channels, the distortion of the proposed approach is comparable to that of COVQ. Simulation confirmed
that our proposed approach is a fast and practical method for VQ over noisy channels.
Figures - uploaded by John Sum
Author content
All figure content in this area was uploaded by John Sum
Content may be subject to copyright.
To deal with the problem that the initialisation method based on random selection may provide a suboptimal codebook of vector quantisation (VQ); an improved method is proposed. In the proposed method, Hadamard transform is performed on training vectors, and then the transformed vectors are sorted according to their first elements. The ordered transformed vectors are partitioned into groups. The initial codebook is composed of the mid vector of each ordered group. This method has a better performance and can be used as the initialisation method of VQ to improve and speed up codebook design.
This chapter discusses the quadratic assignment problems (QAPs). The benefit or cost resulting from an economic activity at some location is dependent on the locations of other facilities. To model such location problems, QAP is introduced. The objective assigns the plants to the possible sites such that the total cost of building and operating the plants becomes minimal. There are many application areas that are modeled by QAPs. The chapter surveys the theory and the solution procedures for QAPs and outlines the various exact and approximate solution methods. There is only one successful exact solution technique available that is based on matrix reductions and on the Gilmore–Lawler bounds. It is demonstrated that the trace form provides a very convenient tool to derive the theory of this method. There is also a second use of the trace formulation. This approach yields completely new competitive bounds for the symmetric case based on the eigenvalues of the underlying matrices. The eigenvalue related bounds give access to an optimal reduction procedure and also help to characterize QAPs that are almost linear. The chapter describes some preliminary numerical comparisons with the classical lower bounds.
An efficient and intuitive algorithm is presented for the design of vector quantizers based either on a known probabilistic model or on a long training sequence of data. The basic properties of the algorithm are discussed and demonstrated by examples. Quite general distortion measures and long blocklengths are allowed, as exemplified by the design of parameter vector quantizers of ten-dimensional vectors arising in Linear Predictive Coded (LPC) speech compression with a complicated distortion measure arising in LPC analysis that does not depend only on the error vector.
From the Publisher:All relevant components of a mobile radio system, from digital modulation techniques over channel coding through to network aspects, are determined by the propagation characteristics of the channel. Therefore, a precise knowledge of mobile radio channels is crucial for the development, evaluation and test of current and future mobile radio communication systems. This volume deals with the modelling, analysis, and simulation of mobile fading channels and provides a fundamental understanding of many issues that are currently being investigated in the area of mobile fading channel modelling. The author strongly emphasises the detailed derivation of the presented channel models and conveys a high degree of mathematical unity to the reader.Introduces the fundamentals of stochastic and deterministic channel modelsFeatures the modelling and simulation of frequency-nonselective fading channels (Rayleigh channels, Rice channels, generalized Rice channels, Nakagami channels, various types of Suzuki channels, classical and modified Loo model)Presents the modelling and simulation of frequency-selective fading channels (WSSUS models, DGUS models, channel models according to COST 207)Discusses the methods used for the design and realization of efficient channel simulatorsExamines the design, realization, and analysis of fast channel simulatorsIncludes MATLAB programs for the evaluation and simulation of mobile fading channelsMATLAB is a registered trademark of The MathWorks, Inc.Telecommunication engineers, computer scientists, and physicists will all find this text both informative and instructive. It is also be an indispensable reference for postgraduate and senior undergraduate students of telecommunication and electrical engineering.
The authors present a polynomial-time codeword-mapping algorithm
which seeks to minimize the Euclidean distance between all pairs of
codewords with a relative Hamming distance of 1, permitting the robust
performance of vector quantizations in noisy channels, without a
sacrifice in coding bandwidth or quantizer performance. The algorithm is
based upon network design considerations, and useful gains of several
decibels can be obtained in the error distortion by using the new
codeword mapping, without a sacrifice in coding bandwidth or quantizer
performance. Results with application to image coding are given, and it
is also shown that some natural codes are good robust codes under
selected conditions
We compare a number of training algorithms for competitive learning networks applied to the problem of vector quantization for data compression. A new competitive-learning algorithm based on the “conscience” learning method is introduced. The performance of competitive learning neural networks and traditional non-neural algorithms for vector quantization is compared. The basic properties of the algorithms are discussed and we present a number of examples that illustrate their use. The new algorithm is shown to be efficient and yields near-optimal results. This algorithm is used to design a vector quantizer for a speech database. We conclude with a discussion of continuing work.
T. Kohonen has described an algorithm for fitting a k-dimensional grid of points to a set of points taken from a k-manifold in Rn, for k⩽n. The algorithm is inspired by a neural model and bears some of the marks of its ancestry. In this paper we show that if the process converges, it converges to a locally 1-1 mapping of the grid onto the manifold. Hitherto this result has only been proved for the case where k=1.
This monograph gives a tutorial treatment of new approaches to
self-organization, adaptation, learning and memory. It is based on
recent research results, both mathematical and computer simulations, and
lends itself to graduate and postgraduate courses in the natural
sciences. The book presents new formalisms of pattern processing:
orthogonal projectors, optimal associative mappings, novelty filters,
subspace methods, feature-sensitive units, and self-organization of
topological maps, with all their computable algorithms. The main
objective is to provide an understanding of the properties of
information representations from a general point of view and of their
use in pattern information processing, as well as an understanding of
many functions of the brain. In the second edition two new chapters on
neural computing and optical associative memories have been added.