15
ADVANCES FOR IN-VEHICLE AND MOBILE SYSTEMS Challenges for International Standards
ADVANCES FOR IN-VEHICLE AND MOBILESYSTEMS
Challenges for International Standards
ADVANCES FOR IN-VEHICLE AND MOBILESYSTEMS
Challenges for International Standards
Edited by
HOseyin AbutSan DiegoState University,San Diego,California, USA and Sabanci University, Turkey<[email protected]>
John H.L. HansenCenter for RobustSpeechSystems(CRSS)Department of ElectricalEngineering,Erik JonssonSchoolof Engineering & ComputerScienceUniversityof Texas at Dallas, Richardson, TX, USA<[email protected]>
Kazuya TakedaDepartment of MediaScienceNagoyaUniversity, Nagoya,Japan<[email protected]>
~ Springer
Hiiseyin AbutSan Diego State UniversitySan Diego, California,USAandSabanci UniversityIstanbul,Turkey
John H.L. HansenCenter for Robust Speech Systems(CRSS)Departmentof ElectricalEngineeringErik Jonsson Schoolof Engineering & ComputerScienceUniversityof Texas at DallasRichardson, TX, USA
KazuyaTakedaDepartmentof Media ScienceNagoya UniversityNagoya, Japan
Advancesfor In-Vehicle and Mobile Systems: Challenges for International Standards
Library of Congress Control Number: 2004051229
ISBN-IO 0-387-33503-XISBN-13 978-0-387-33503-2
Printed on acid-free paper.
e-ISBN-IO 0-387-45976-6e-ISBN-13 978-0-387-45976-9
© 2007SpringerSeienee+Business Media, LLCAll rights reserved. This work may not be translated or copied in whole or in part without the writtenpermission of the publisher (SpringerSeienee+Business Media, LLC,233 Spring Street,New York, NY10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Usc inconnection with any form of information storageand retrieval, electronic adaptation, computersoftware,or by similaror dissimilarmethodology now knowor hereafterdeveloped is forbidden.
The usc in this publication of trade names, trademarks, service marksand similar terms,even if they arenot identified as such, is not to be takenas an expression of opinion as to whetheror not they arc subjectto proprietary rights.
9 8 7 6 5 4 3 2
springer.com
Foreword
The past, the present, and a glimpse of the future of the use ofDigital Signal Processing (DSP) in vehicles are contained in the pages of thistextbook. The papers within its covers present the results of an impressivearray of research built on a solid base of 40 years of signal processing andautomatic speech recognition. You will read papers that push parameters ,tease out nuances, and point the reader into new directions as they addressvarious aspects of the complicated man-machine interface that occurs in amoving vehicle requiring extensive cognitive demands. It can be a dauntinginterface, one that encompasses a challenging and changing acousticalenvironment with shifting situational wants and needs for drivers.
The past of DSP is clearly here. Prior research is often cited, and aknowledgeable reader will also see it in the starting points, assumptions, andtechniques employed. The present of DSP is also certainly here. Whatbegan with few people, limited resources, and little public attention hasmushroomed extensively in a relatively short period. The advent of suchtechnologies as cellular telephony and accurate, interactive GPS navigationalsystems has made the public aware of the possibilities of DSP, and withpublic knowledge has come public demand. So, what had been the interestand passion of a few is now the pursuit of many. Public demand for voiceactivated systems for in-vehicle environments for the sake of comfort andconvenience is growing exponentially. The research presented in this bookis intended to meet the public's demand in an effective, unobtrusive , andresponsible way. Some of the topics covered include reduction, suppression,and control of noise in vehicles to enhance speech discrimination; speechrecognition systems to facilitate speech control-oriented tasks; biometricrecognition systems to identify individual drivers; and provide dialogue
VI Foreword
management based on driver workload. Each research area discussed facesunique technical challenges in its development and implementation.
What's more, each research area also faces a special challengebecause human beings are involved . As these papers show, the science hasclearly advanced, but the capabilities of human users have not. For example,think about the complexity of developing a speech recognition system toenhance telematics use and how a driver might employ that system. Thenthink about the relatively simple, basic driving task. We know it is relativelysimple by virtue of the sheer number of people who can successfully drive avehicle. Millions, from teenagers to grandparents, drive. The availability ofvoice-activated telematics may well enhance the comfort, convenience, and,in some circumstances, the safety of drivers while they are carrying out thisbasic task. Still, we also have millions of accidents and tens of thousands ofdeaths each year, suggesting that the driving task is not always a simple one.When it is not, when the workload is high and time window is short, manydrivers are ill-equipped to meet these changing and increased demands .What role does the voice-activated telematics system have at those times?
The answer to that question brings us to the future of DSP.Technologically, solutions are evolving rapidly , thanks to the bright men andwomen reporting on their research in this book. Future design anddevelopment will lead to a seamless implementation of hands-freeinteraction among digital devices in vehicles . In addition, systems will notonly meet the wants and needs of drivers but also accommodate theirfrailties by factoring in the workload context. Safety demands require, andeach researcher and system designer has to be ever mindful that, to besuccessful, these efforts must first do no harm.
So, read on, learn, be impressed, and see the future.
Bruce A. Magladry, DirectorOffice of Highway SafetyNational Transportation Safety Board, USA
Contents
Foreword
Contents
Contributing Authors
Introduction
Chapter 1Experiments on Decision Fusion for Driver RecognitionHakan Erdogan, Aytiil Ercil and Hiiseyin Abut
v
VB
Xl
XV
Chapter 2Driver Recognition System Using FNN and Statistical Methods 11Abdul Wahab, Tan Chin Keong , Hiiseyin Abut and Kazuya Takeda
Chapter 3Driver Identification Based on Spectral Analysis of DrivingBehavioral Signals 25Yoshihiro Nishiwaki, Koji Ozawa, Toshihiro Wakita ,Chiyomi Miyajima, Katsunobu Itou, and Kazuya Takeda
Chapter 4An Artificial-Vision Based Environment Perception System 35s. Nogueira, Y. Ruichek, F. Gechter, A. Koukam, and F. Charpillet
Vlll Contents
Chapter 5Variable Time-Scale Multimedia Streaming Over 802.11Inter-Vehicle Ad-hoc Networks 47Antonio Servetti, Enrico Masala, Paolo Bucciol,and Juan Carlos De Martin
Chapter 6A Configurable Distributed Speech Recognition System 59HaitianXu, Zheng-Hua Tan, Paul Dalsgaard,Ralf Mattethat, and Berge Lindberg
Chapter 7Embedded MobilePhoneDigit-Recognition 71Christophe Levy, Georges Linares, PascalNocera,and Jean-Francois Bonastre
Chapter 8On The Complexity-Performance Tradeoffof Two ActiveNoiseControlSystems for Vehicles 85Pedro Ramos, Luis Vicente, Roberto Torrubia, Ana Lopez,Ana Salinas, and Enrique Masgrau
Chapter 9Comparative Studies on Single-Channel De-Noising Schemesfor In-carSpeech Enhancement 97Weifeng Li, Katunobu Itou, KazuyaTakeda, and FumitadaItakura
Chapter 10Advances in Acoustic Noise Tracking for RobustIn-vehicle Speech Systems 109MuratAkbacak and John H.L. Hansen
Chapter 11SpeakerSourceLocalization Using Audio-Visual Dataand ArrayProcessing Based SpeechEnhancement forIn-vehicle Environments 123Xianxian Zhang, John H.L. Hansen, KazuyaTakeda,Toshiki Maeno, Kathryn Arehart
Contents
Chapter 12Estimation of Active Speaker's Direction Using ParticleFilters for In-vehicle EnvironmentMitsunori Mizumachi and Katsuyuki Niyada
Chapter 13Noise Reduction Based on Microphone Array andPost-Filtering for Robust Speech Recognition in CarEnvironmentsJunfeng Li and Masato Akagi
Chapter 14ICA-Based Technique in Air and Bone-ConductiveMicrophones for Speech EnhancementZhipeng Zhang, Kei Kikuiri, Nobuhiko Naka, and Tomoyuki Ohya
Chapter 15Acoustic Echo Reduction in a Two-Channel SpeechReinforcement System for VehiclesAlfonso Ortega, Eduardo Lleida, Enrique Masgrau, Luis Buera,
and Antonio Miguel
Chapter 16Noise Source Contribution of Accelerating Cars andSubjective EvaluationsShunsuke Ishimitsu
Chapter 17Study on Effect of Speaker Variability and DrivingConditions on the Performance of an ASR EngineInside a VehicleShubha Kadambe
Chapter 18Towards Robust Spoken Dialogue Systems UsingLarge-Scale In-car Speech CorpusYukiko Yamaguchi, Keita Hayashi, Takahiro Ono,
Shingo Kato, Yuki Irie, Tomohiro Ohno, Hiroya Murao,
Shigeki Matsubara, Nobuo Kawaguchi, Kazuya Takeda
141
153
167
177
189
201
211
IX
x
Chapter 19Exploitation of Context Information for NaturalSpeech Dialogue Management in Car EnvironmentsMarkus AblAf3meier and Gerhard Rigoll
Chapter 20Cross Platform Solution of Communication andVoice/ Graphical User Interface for MobileDevicesin VehiclesGeza Nemeth, Geza Kiss, BalintToth
Contents
223
237
Chapter 21A Studyof Dialogue Management Principles Correspondingto The Driver's Workload 251MakotoShioya, Takuya Nishimoto, Juhei Takahashi,and Hideharu Daigo
Chapter 22Robust Multimodal DialogManagement for MobileEnvironments 265Jeonwoo Ko, Fumihiko Murase, Teruko Mitamura,Eric Nyberg, Nobuo Hataoka, HirohikoSagawa, Yasunari Obuchi,Masahiko Tateishi, and IchiroAkahori
Index 279
Contributing Authors
Abdul Wahab, Nanyang Technological University, Singapore
Markus AblaOmeier, Munich University of Technology, Germany
HUseyin Abut, San Diego State University, USA and Sabanci University,Turkey
Masato Akagi, Advanced Institute of Science and Technology, Japan
Ichiro Akahori, Denso Corporation, Japan
Murat Akbacak, University of Texas at Dallas, USA
Kathryn Arehart, University of Colorado at Boulder, USA
Jean-Francais Bonastre, Laboratoire Inforrnatique Avignon , France
Paolo Bucciol, Politecnico di Torino, Italy
Luis Buera , University of Zaragoza. Spain
Francois Charpillet, Systems and Transportation Laboratory, France
Hideharu Daigo, Automobile Research Institute, Japan
Paul Dalsgaard, Aalborg University, Denmark
Juan Carlos De Martin , Politecnico di Torino, Italy
AytUl Erell, Sabanci University, Turkey
Hakan Erdogan, Sabanci University, Turkey
Franck Gechter, Systems and Transportation Laboratory, France
Xli Contributing Authors
John H. L. Hansen, University of Texas at Dallas, USA
Nobuo Hataoka, Hitachi Advanced Research Laboratory, Japan
Keita Hayashi, Nagoya University, Japan
Yuki Irie, Nagoya University, Japan
Shunsuke Ishimitsu, University of Hyogo, Japan
Fumitada Itakura , Meijo University, Japan
Katsunobu Itou, Nagoya University, Japan
Shubha Kadambe, Signal and Image Processing, Office of NavalResearch/University of Maryland, College Park, USA
Shingo Kato, Nagoya University, Japan
Nobuo Kawaguchi, Nagoya University, Japan
Kei Kikuiri, NTT DoCoMo Multimedia Laboratories, Japan
Geza Kiss, Budapest University of Technology and Economics, Hungary
Jeonwoo Ko, Carnegie Mellon University , USA
Abderrafiaa Koukam, Systems and Transportation Laboratory, France
Eduardo Lleida, University of Zaragoza. Spain
Christophe Levy, Laboratoire Informatique Avignon, France
Junfeng Li, Advanced Institute of Science and Technology, Japan
Weifeng Li, Nagoya University, Japan
Georges Linares, Laboratoire Informatique Avignon, France
Ana Lopez, University of Zaragoza. Spain
Borge Lindberg, Aalborg University, Denmark
Toshiki Maeno, Nagoya University, Japan
Bruce A. Magladry, Office of Highway Safety, National TransportationSafety Board, USA
Enrico Masala, Politecnico di Torino, Italy
Enrique Masgrau, University of Zaragoza. Spain
Antonio Miguel, University of Zaragoza. Spain
Contributing Authors X1ll
Ralf Mathetat, Arhus Technology Institute, Denmark
Shigeki Matsubara, Nagoya University, Japan
Teruko Mitamura, Carnegie Mellon University, USA
Chiyomi Miyajima, Nagoya University, Japan
Mitsunori Mizumachi, Kyushu Institute of Technology, Japan
Hiroya Murao, SANYO Electric Company, Japan
Fumihiko Murase, Carnegie Mellon University, USA
Nobuhiko Naka, NTT DoCoMo Multimedia Laboratories, Japan
Geza Nemeth, Budapest University of Technology and Economics ,Hungary
Takuya Nishimoto, The University of Tokyo, Japan
Yoshihiro Nishiwaki, Nagoya University, Japan
Katsuyuki Niyada, Kyushu Institute of Technology , Japan
Pascal Nocera, Laboratoire Informatique Avignon, France
Sergio Nogueira, Systems and Transportation Laboratory, France
Eric Nyberg, Carnegie Mellon University, USA
Tomohiro Ohno, Nagoya University, Japan
Tomoyuki Ohya, NTT DoCoMo Multimedia Laboratories, Japan
Takahiro Ono, Nagoya University, Japan
Alfonso Ortega, University of Zaragoza. Spain
Koji Ozawa, Nagoya University, Japan
Pedro Ramos, University of Zaragoza. Spain
Gerhard Rigoll, Munich University of Technology, Germany
Yassine Ruichek, Systems and Transportation Laboratory, France
Hirohiko Sagawa, Hitachi Advanced Research Laboratory, Japan
Ana Salinas, University of Zaragoza. Spain
Antonio Servetti, Politecnico di Torino, Italy
Makoto Shioya, Hitachi Systems Development Laboratory, Japan
XIV Contributing Authors
Juhei Takahashi, Automobile Research Institute, Japan
Kazuya Takeda,Nagoya University, Japan
Chin KeongTan, Nanyang Technological University, Singapore
Zheng-Hua Tan, Aalborg University, Denmark
Masahiko Tateishi, Denso Corporation, Japan
Balint Toth, Budapest University of Technology and Economics, Hungary
Roberto Torrubia, University of Zaragoza. Spain
Luis Vicente, University of Zaragoza. Spain
Toshihiro Wakita,Toyota Central R&D Laboratories, Japan
HaitanXu, Aalborg University, Denmark
YukikoYamaguchi, Nagoya University, Japan
Xianxian Zhang, University of Texas at Dallas, USA
ZhipengZhang, NTT DoCoMo Multimedia Laboratories, Japan
Introduction
In September 2005, the "Second Biennial Workshop on DSP (digitalsignal processing) for Mobile and Vehicular Systems" took place inSesimbra, Portugal with 32 excellent papers present from all over the world,with a Keynote Address entitled "Information Overload and DriverDistraction: The Road to Disaster," delivered by Bruce A. Magladry,Director of the Office of Highway Safety, U.S. National TransportationSafety Board (NTSB), and a panel discussion with experts from academic,industry and federal agencies. This meeting represented a continuation fromthe first workshop in Nagoya, Japan, April 2003, and this book reflects theoffspring of the 2005 workshop. After carefully reviewing all papers, 22presentations from the workshop were selected and authors were asked toformulate extended book chapters of their original papers in order to providea broad coverage of the fields in DSP for Mobile and Vehicular Systems,namely, Driver and Driving Environment Recognition, TelecommunicationApplications, Noise Reduction and Dialog Systems for In-Vehicle systems.The chapters contained are therefore naturally categorized into four parts.
In the first part, a new and an emerging research field, Driver andDriving Environment Recognition is introduced and addressed with fourcomplementary chapters. These chapters report on the application of signalprocessing technologies to characterizing human behavior while driving andencouraging further research efforts on the analysis of driver behavior.
The second part consists of three chapters dedicated to the mostimportant application in Mobile and Vehicular Systems, (i.e.,Telecommunication Applications). This part considers technologies forwireless communication systems, a distributed speech recognition paradigm,and embedded platforms.
XVI Introduction
The third and largest part of this book addresses a major challenge forMobile and Vehicular Systems, namely Noise Reduction , which has everchanging environmental noise that seriously degrades performance of thesesystems. Various research efforts that range from measuring and controllingthe acoustic noise to an approach for audio-visual data fusion are discussed.ICA, Particle Filtering and other state of the art methodologies are alsostudied in this third part.
The topic of the last part of the book, Dialog Systems for In-Vehicle ,is an essential issue for not only interface efficiency , but also for safety andcomfort while driving. In this last part, six chapters discuss topics such asinteraction between driving and dialogue, corpus, bimodal interface anddialogue strategies .
We hope this book will provide an up to date treatment of Mobile andVehicular Systems, with new ideas for researchers and comprehensive set ofreferences for engineers in related fields. We thank all those whoparticipated in the 2005 workshop, we acknowledge support from the U.S.National Science Foundation and Nagoya University for their support inorganizing the Biennial DSP for In-Vehicle and Mobile Systems inSesimbra, Portugal, Sept. 2-3, 2005. We thank Dr. Pongtep Angkititrakulfrom CRSS-UTD for his assistance in book formatting/editing, and we wishto express our appreciation to Springer Publishing for ensuring a smooth andefficient publication process for this textbook.
The Editors. Hiiseyin Abut. John H.L. Hansen, Kazuya Takeda
