Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Feature Extraction for Low-Frequency Signal Classification
KTH, School of Engineering Sciences (SCI), Physics.
2008 (English)Doctoral thesis, comprehensive summary (Other scientific)
Abstract [en]

The development of the International Monitoring System (IMS) for the verification of the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO) has led to a rapid revival of interest in infrasound. Furthermore, the installation of low-frequency sensors at seismic sites has increased in recent years, providing researchers with large and heterogeneous data-sets in near real-time. New techniques are needed to better process all of this data and to extract meaningful information quickly for various applications. In particular, there is a need to find distinct features in the infrasonic signals that allow one to distinguish low level nuclear tests from seismic events.

In this thesis three methods for feature extraction from infrasound and other types of low frequency signal data are discussed: (1) discrete wavelets transforms (DWTs); (2) time scale spectra (TSSs) using continuous wavelet transforms (CWTs); and (3) empirical mode decomposition (EMD). The dimensionality of the feature space can range from a few to thousands. For processing highdimensional data we use multi-modal data space to find low-dimensional structures. The advantage of this space is that multiple metrics of similarity are converted into one single Euclidean space.

The overall goal of our research is a system for automatic identification and classification of lowfrequency signals in real-time that is easy to implement in hardware. In this thesis we present our design and implementation of the discrete wavelet transform (DWT) on FPGAs for processing a continuous flow of data to obtain optimal extraction of feature information. FPGA simulation and implementation has been realized by using the polyphase structure, the filter reuse method and techniques such as pipelining and basic operations on the QUARTUS II platform. VHDL has been used to describe the functionality of the discrete wavelet transform and ModelSim has been used for the functional verification.

Advancements in electronics provide a vital new option for implementation of low-frequency smart sensors that can perform signal processing close to the sensors and transmit the data wirelessly. These smart sensors can improve the efficiency of an automatic classification system and reduce the cost of actual infrasound microphones. The design of a digital wireless data acquisition system using a QF4512 programmable signal converter from Quickfilter Technologies, a MSP430 microcontroller from Texas Instruments and a F2M03GLA Bluetooth module from Free2move for infrasonic records is also presented in this thesis. The digital wireless data acquisition system has passed extensive laboratory and field tests (e.g. with man-made explosions).

A study of using a miniature 3-axis fluxgate magnetometer to get fingerprints from ferrous objects is also presented in this thesis. In this experiment, distinguishing features of iron samples of four different shapes were determined using wavelet methods. Systematic differences were observed between the signatures of the four shaped iron samples.

Place, publisher, year, edition, pages
Stockholm: KTH , 2008. , viii, 61 p.
Series
Trita-FYS, ISSN 0280-316X ; 2008:2
Keyword [en]
infrasound, seismic signals, feature extraction, wavelets, fingerprints, mining, multimodal data space
National Category
Atom and Molecular Physics and Optics
Identifiers
URN: urn:nbn:se:kth:diva-4661ISBN: 978-91-7178-856-6 (print)OAI: oai:DiVA.org:kth-4661DiVA: diva2:13300
Public defence
2008-03-31, FA 31, AlbaNova Universitetscentrum, Roslagstullsbacken 21, Stockholm, 10:00
Opponent
Supervisors
Note
QC 20100713Available from: 2008-03-06 Created: 2008-03-06 Last updated: 2010-07-13Bibliographically approved
List of papers
1. Classification of infrasound events with various machine learning techniques
Open this publication in new window or tab >>Classification of infrasound events with various machine learning techniques
2007 (English)In: CITSA 2007 - Int. Conference on Cybernetics and Information Technologies, Systems and Applications and CCCT 2007 - Int. Conference on Computing, Communications and Control Technologies, Proceedings: Vol II / [ed] Savoie M; Aguilar J; Chu HW; Zinn CD; AroshaS SMN, International Institute of Informatics and Systemics, 2007, 191-195 p.Conference paper, Published paper (Refereed)
Abstract [en]

This paper presents classification results for infrasonic events using practically all well-known machine learning algorithms together with wavelet transforms for preprocessing. We show that there are great differences between different groups of classification algorithms and that nearest neighbor classifiers are superior to all others for accurate classification of infrasonic events.

Place, publisher, year, edition, pages
International Institute of Informatics and Systemics, 2007
Keyword
classification, machine lcarning, pattern recognition, wavelets
National Category
Engineering and Technology
Identifiers
urn:nbn:se:kth:diva-14098 (URN)000254391300036 ()2-s2.0-84890841348 (Scopus ID)1934272086 (ISBN)978-193427208-4 (ISBN)
Conference
4th International Conference on Cybernetics and Information Technologies, Systems and Applications, CITSA 2007, Jointly with the 5th International Conference on Computing, Communications and Control Technologies, CCCT 2007; Orlando, FL; United States; 12 July 2007 through 15 July 2007
Note

QC 20100713

Available from: 2010-07-13 Created: 2010-07-13 Last updated: 2014-02-17Bibliographically approved
2. Comparison of three feature extraction techniques to distinguish between different infrasound signals
Open this publication in new window or tab >>Comparison of three feature extraction techniques to distinguish between different infrasound signals
2007 (English)In: Progress in Pattern Recognition / [ed] Singh S; Singh M, 2007, 75-82 p.Conference paper, Published paper (Refereed)
Abstract [en]

The main aim of this paper is to compare three feature extraction techniques, Discrete Wavelet Transform, Time Scale Spectrum using Continuous Wavelet Transforms, and Cepstral Coefficients and their derivatives, for the purposes of classifying time series type ;signal data. The features are classified by two types of neural networks. The paper draws a number of important conclusions on the suitability of these features for analysis, and provides a good comparative evaluation on four different data sets.

Series
Advances in pattern recognition (series), ISSN 1617-7916
National Category
Engineering and Technology
Identifiers
urn:nbn:se:kth:diva-14100 (URN)10.1007/978-1-84628-945-3_8 (DOI)000250406000008 ()978-1-84628-944-6 (ISBN)
Conference
International Workshop on Advances in Pattern Recognition,Loughborough Univ, Loughborough, ENGLAND, 2007
Available from: 2010-07-13 Created: 2010-07-13 Last updated: 2011-09-13Bibliographically approved
3. Discrimination of nuclear explosions sites by seismic signals using intrinsic mode functions and multi-modal data space
Open this publication in new window or tab >>Discrimination of nuclear explosions sites by seismic signals using intrinsic mode functions and multi-modal data space
2008 (English)In: 2008 IEEE International Geoscience and Remote Sensing Symposium - Proceedings; Boston, MA; 6 July 2008 through 11 July 2008, 2008, II895-II898 p.Conference paper, Published paper (Refereed)
Abstract [en]

Signal processing and feature extraction are investigated using the Empirical Mode Decomposition (EMD). It is believed that this approach is well suited for non-linear and non-stationary data. With EMD any complicated set of data can be decomposed into a finite, and usually small number, of functions called Intrinsic Mode Functions (IMFs). A new discriminating system is presented here that is capable of discriminating between different seismic signals from nuclear testing sites based on the IMFs and the multi-modal data space. The advantage of this space is that multiple metrics of similarity are converted into one single Euclidean space. This space is capable of extracting similarities among several signals through a combination of multiple metrics. This is a new way of associating data. After illustrating the technique with an investigation of an audio data example (piano), we examine the characteristics of seismic signals from nuclear testing (explosions). The results presented in this paper indicate that a relatively simple discriminating system can successfully cluster and classify seismic events.

Series
International Geoscience and Remote Sensing Symposium (IGARSS), Volume 2, Issue 1, 2008
Keyword
Hilbert-Huang transform; Intrinsic mode function; Multi-modal data space; Nuclear testing; Seismic data
National Category
Engineering and Technology
Identifiers
urn:nbn:se:kth:diva-14103 (URN)10.1109/IGARSS.2008.4779139 (DOI)2-s2.0-66549094935 (Scopus ID)
Note
QC 20100713Available from: 2010-07-13 Created: 2010-07-13 Last updated: 2010-07-13Bibliographically approved
4. Hardware implementation of 1D wavelet transform on an FPGA for infrasound signal classification
Open this publication in new window or tab >>Hardware implementation of 1D wavelet transform on an FPGA for infrasound signal classification
2008 (English)In: IEEE Transactions on Nuclear Science, ISSN 0018-9499, E-ISSN 1558-1578, Vol. 55, no 1, 9-13 p.Article in journal (Refereed) Published
Abstract [en]

Infrasound is a low frequency acoustic phenomenon that typically ranges from 0.01 to 20 Hz. The data collected from infrasound microphones are presented online by the infrasound monitoring system operating in Northern Europe, i.e., the Swedish-Finnish Infrasound Network (SFIN). Processing the continuous flow of data to extract optimal feature information is important for real-time signal classification. Performing wavelet decomposition on the real-time signals is an alternative. The purpose of this paper is to present the design and FPGA implementation of discrete wavelet transforms (DWT) for real-time infrasound data processing; our approach uses only two FIR filters, a high-pass and a low-pass filter. A compact implementation was realized with pipelining techniques and multiple use of generalized building blocks. The design was described in VHDL and the FPGA implementation and simulation were performed on the QUARTUS II platform.

Keyword
discrete wavelet transform (DWT), FPGA, filtering, infrasound
National Category
Engineering and Technology
Identifiers
urn:nbn:se:kth:diva-14105 (URN)10.1109/TNS.2007.914322 (DOI)000253224700003 ()2-s2.0-39049131227 (Scopus ID)
Note
QC 20100713Available from: 2010-07-13 Created: 2010-07-13 Last updated: 2017-12-12Bibliographically approved
5. Filtering and extracting features from infrasound data
Open this publication in new window or tab >>Filtering and extracting features from infrasound data
2005 (English)In: 2005 14TH IEEE-NPSS Real Time Conference: Stockholm; 4 June 2005 through 10 June 2005, 2005, 451-455 p.Conference paper, Published paper (Refereed)
Abstract [en]

There are many reasons for using infrasound, i.e. low frequency sound, to monitor various events. Inherent features like its long-distance propagation and the use of simple, ground based equipment in very flexible system are some. The disadvantage is that it is a slow system due to the speed of sound. In this papr we try to show that there are several other advantages if one can extract all the features of the signal. In this way it is hoped that we can get a fingerprint of the event that caused the infrasound. Rayleigh waves and sound from epicentre may be obtained for earthquakes, pressure pulses and electro jets from aurora, core radius and funnel shape from tornados, etc. All these possibilities are suggestive for further R&D of the infrasound detection systems.

National Category
Engineering and Technology
Identifiers
urn:nbn:se:kth:diva-14106 (URN)10.1109/RTC.2005.1547494 (DOI)2-s2.0-33751438758 (Scopus ID)978-0-7803-9183-3 (ISBN)
Note
QC 20100713Available from: 2010-07-13 Created: 2010-07-13 Last updated: 2010-11-10Bibliographically approved
6. Obtaining "images" from iron objects using a 3-axis fluxgate magnetometer
Open this publication in new window or tab >>Obtaining "images" from iron objects using a 3-axis fluxgate magnetometer
Show others...
2007 (English)In: Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, ISSN 0168-9002, E-ISSN 1872-9576, Vol. 580, no 2, 1105-1109 p.Article in journal (Refereed) Published
Abstract [en]

Magnetic objects can cause local variations in the Earth's magnetic field that can be measured with a magnetometer. Here we used triaxial magnetometer measurements and an analysis method employing wavelet techniques to determine the "signature" or "fingerprint" of different iron objects. Clear distinctions among the iron samples were observed. The time-dependent changes in the frequency powers were extracted by use of the Morlet wavelet corresponding to frequency bands from 0.1 to 100 Hz. (c) 2007 Elsevier B.V. All rights reserved.

Keyword
magnetometer, feature extraction, wavelets, fingerprints
National Category
Physical Sciences
Identifiers
urn:nbn:se:kth:diva-14107 (URN)10.1016/j.nima.2007.06.070 (DOI)000250128000058 ()2-s2.0-34548476793 (Scopus ID)
Note
QC 20100713Available from: 2010-07-13 Created: 2010-07-13 Last updated: 2017-12-12Bibliographically approved
7. Data acquisition and digital filtering for infrasonic records on active volcanoes
Open this publication in new window or tab >>Data acquisition and digital filtering for infrasonic records on active volcanoes
2007 (English)In: Sensors & Transducers Journal, ISSN 2306-8515, E-ISSN 1726-5479, ISSN 1726-5479, Vol. 77, no 3, 1058-1064 p.Article in journal (Refereed) Published
Abstract [en]

This paper presents the design of a digital data acquisition system for volcanic infrasound records. The system includes four electret condenser element microphones, a QF4A512 programmable signal converter from Quickfilter Technologies and a MSP430 microcontroller from Texas Instruments. The signal output of every microphone is converted to digital via a 16-bit Analog to Digital Converter (ADC). To prevent errors in the conversion process, Anti-Aliasing Filters are employed prior to the ADC. Digital filtering is performed after the ADC using a Digital Signal Processor, which is implemented on the QF4A512. The four digital signals are summed to get only one signal. Data storing and digital wireless data transmission will be described in a future paper.

Keyword
Volcanic Infrasound, Data Acquisition, Digital Filtering, Digital Signal Processor
National Category
Engineering and Technology
Identifiers
urn:nbn:se:kth:diva-14109 (URN)
Note

QC 20100713

Available from: 2010-07-13 Created: 2010-07-13 Last updated: 2017-12-12Bibliographically approved

Open Access in DiVA

No full text

Search in DiVA

By author/editor
Chilo, José
By organisation
Physics
Atom and Molecular Physics and Optics

Search outside of DiVA

GoogleGoogle Scholar

isbn
urn-nbn

Altmetric score

isbn
urn-nbn
Total: 2832 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf