ServicenavigationHauptnavigationTrailKarteikarten


Forschungsstelle
EU FRP
Projektnummer
95.0345a
Projekttitel
INSPECT: Innovative signal processing exploiting chaotic dynamics
Projekttitel Englisch
INSPECT: Innovative signal processing exploiting chaotic dynamics

Texte zu diesem Projekt

 DeutschFranzösischItalienischEnglisch
Schlüsselwörter
-
-
-
Anzeigen
Alternative Projektnummern
-
-
-
Anzeigen
Forschungsprogramme
-
-
-
Anzeigen
Kurzbeschreibung
-
-
-
Anzeigen
Partner und Internationale Organisationen
-
-
-
Anzeigen
Abstract
-
-
-
Anzeigen
Datenbankreferenzen
-
-
-
Anzeigen

Erfasste Texte


KategorieText
Schlüsselwörter
(Englisch)
Chaos; communications; watermarking; channel noiuse; copyreight protection; differential chaos shift keying; public key
Alternative Projektnummern
(Englisch)
EU project number: 31103
Forschungsprogramme
(Englisch)
EU-programme: 4. Frame Research Programme - 1.3 Telematic systems
Kurzbeschreibung
(Englisch)
See abstract
Partner und Internationale Organisationen
(Englisch)
Coordinator: University College Dublin (UK)
Abstract
(Englisch)
The aim of the project was to explore the possibilities of using signals produced by chaotic systems for communications and for watermarking.
Overall, the project has lead to a functioning prototype of a communication system whose core uses the method of differential chaos shift keying. It comprises the baseband chaos generator, the circuitry for the modulation/demodulation and amplification and the antennas, as well as the realization of the algorithm that extracts the information on the receiver side. Its performance is close to competitive with current technology. In addition, theoretical results on performance with respect to noisy channels, and the corresponding optimisation, were obtained.
The project has also lead to the development of different fairly robust watermarking techniques, both for image and audio documents. As a new theoretical approach, the watermarking system was interpreted as a communication system, which allowed to carry over spread spectrum notions and methods from communications to watermarking.
The main contributions of EPFL in the project were the following:
For the problem of communications with chaos, two approaches for the decoding of the noise perturbed chaotic signal in order to extract the information on the receiver side were investigated. The first starts with a noise-removal algorithm that is followed by the information extraction, whereas the second performs the information extraction directly. It turned out that the first method is not efficient, because the noise-removal performance criterion is not adapted to the information extraction. As far as the direct method is concerned, the optimal receiver algorithm was defined and developed. Optimality means here lowest bit error rate for a given signal to noise ratio, normalized with the energy of the signal used to send one bit. The optimal differential chaos shift keying receiver outerforms the conventional correlation receiver. However, since it is computationally too complex to be implemented for large spreading factors, a suboptimal method based on a Hidden Markov Model associated with the chaotic system and the Viterbi algorithm was also developed.
As far as watermarking is concerned, a simple time-domain based system that can be used for authentication was developed. A more elaborate and robust system where the watermark is added in the transform domain was developed based on the interpretation as a communication system. It can be used for copyright protection of image and audio documents. An improvement using perceptual thresholds was shown to be vulnerable to specific attacks if it is not carefully designed. In this context, a new attack, the mask attack, was developed. Finally, it was shown that chaos can be used to design a 'public key' watermarking system. Here, the knowledge of the chaotic system that has produced the watermark allows to detect it, but not to remove it. Thus, this knowledge can be divulgated as a 'public key'. It must be added, however, that a generic attack can make the watermark invisible to those who only possess the public key.
Datenbankreferenzen
(Englisch)
Swiss Database: Euro-DB of the
State Secretariat for Education and Research
Hallwylstrasse 4
CH-3003 Berne, Switzerland
Tel. +41 31 322 74 82
Swiss Project-Number: 95.0345a