244 Pages
English

Analysis and modeling of driver behavior for assistance systems at road intersections [Elektronische Ressource] / Marina Plavsic

-

Gain access to the library to view online
Learn more

Subjects

Informations

Published by
Published 01 January 2010
Reads 87
Language English
Document size 28 MB

Lehrstuhl fu¨r Ergonomie
der Technischen Universit¨at Munchen¨
Analysis and Modeling
of Driver Behavior for Assistance Systems
at Road Intersections
Marina Plavˇsi´c
Vollstandiger¨ Abdruck der von der Fakultat¨ fur¨ Maschinenwesen der Technischen Universitat¨ Mun-¨
chen zur Erlangung des akademischen Grades eines
Doktor-Ingenieurs (Dr.-Ing.)
genehmigten Dissertation.
Vorsitzender: Univ.-Prof. Dr.-Ing. Markus Lienkamp
Pruf¨ er der Dissertation:
1. Univ.-Prof. Dr. rer. nat. Heiner Bubb, i.R.
2. Univ.-Prof. Dr. phil. Klaus Bengler
3. Univ.-Prof. Gudrun J. Klinker, Ph.D.
DieDissertationwurdeam28.4.2010beiderTechnischenUniversitat¨ Munchen¨ eingereichtunddurch
die Fakultat¨ fur¨ Maschinenwesen am 4.10.2010 angenommen.Mojoj porodiciAcknowledgements
Asayingsays”Agoodworkisneverfinished”,andeventhoughIamhighlymotivatedtoworkfurther
on the presented topic, an end for the dissertation has to be proclaimed at some moment. And this
moment hascome, after threeyears ofgreat time IspentatLehrstuhl fur¨ ErgonomieandFachgebiet
fur¨ Augmented Reality in Munich.
It is a pleasure to express here my great thanks to all who made my work possible. I would like to
thank to DAAD who provided me the financial support and the possibility to continue my education
in Germany. I want to express great appreciation and thanks to my Professors, Prof. Dr. Heiner
Bubb and Professor Ph.D. Gudrun Klinker for accepting to be my supervisors and for believing in the
success of my work from the very beginning. Thank you for your patience in writing the recommen-
dations letters for the prolongation of my scholarship and for always being there to answer on my
questions and doubts. I also want to thank to Prof. Dr. Klaus Bengler who also accepted to be my
supervisor and took a lot of time to give me great input for both, my current and future work.
Coming from the field of Electrical Engineering, the work in Human-Machine Interaction field in Ger-
many presented a big challenge and a valuable experience for me and I would like to thank to all
of you who supported me on my way. I want to thank to my colleges from Lehrstuhl fur¨ Ergonomie
and Fachgebiet fur¨ Augmented Reality, who were helping me to adjust to a new country, teaching
me bayerisch and making me feel as though I really understand it. Special thanks goes to my office
college Marcus Tonnis¨ , who introduced me to both institutes and my office college Darya Popiv who
gave me support in the most critical moments. For the correction of my work I would like to express
my gratefulness to Aleksandar Mihajlovic who was not lazy to read through my whole thesis and
had patience to correct my articles. Thank you Dejan Pangercic, Marko Mihailovic, Nikola Knezevic,
Marcus Tonnis¨ for the feedback on my last draft.
But this thesis would not have been possible without my family who was always there for me and my
sisterDraganawhocametoMunichtosupportmewhenevershefeltthatIcouldneedsupport. Ialso
want to thank to all my friends for their understanding.
And thank you Miki for your patience and confidence and for supporting me during the most difficult
time of writing my thesis.Zusammenfassung
Diese Arbeit befasst sich mit Verkehrssituationen an Kreuzungen, da diese ein hohes Potenzial fur¨
dieStar¨ kungderVerkehrssicherheitundMobilitat¨ bieten. EinestetigeZunahmedesinnerstadtischen¨
Verkehrs,eineimmeralter¨ werdendeGesellschaftundeinschonausgeschopftes¨ Potentialfur¨ Fahreras-
¨ ¨sistenzsysteme fur¨ den Langsverkehr erhohen den Druck auf die Entwicklung von Methoden und
Technologienfur¨ Kreuzungsassistenzsysteme. DiebisherigeAnsatz¨ efur¨ dasDesignvonKreuzungsas-
sistenzsystemen sind entweder durch die fehlende Nutzerakzeptanz, Ineffizienz oder hohe Investi-
tionskosten charakterisiert.
Es wird ein innovativer und ergonomischer Ansatz fur¨ die Fehlervermeidung im Verkehrsgeschehen
¨prasentiert. Es wird gezeigt, dass die Analyse der Fahraufgabe und die Modellierung der Kognition
dieGrundlagefur¨ dieEntwicklungergonomischerAssistenzsystemesind. MitHilfederdurchgefuhr¨ ten
AnalyseninKombinationmitdemStandderForschungwirdeinkognitivesundsimulativesFahrermodel
aufgestellt. Das Ziel dieses Modells ist die Unterstutzung¨ bei der Entwicklung neuer Fahrerassisten-
zssyteme. Desweiteren wurden die Anforderungen an Kreuzungsassistenzsysteme aus den Ergeb-
nissen diverser Fahrsimulatorexperimente abgeleitet. Diese Arbeit zeigt, dass allein die Nutzung
von im Fahrzeug vorhandenen Sensoren und Kartenmeterial die notigen¨ Informationen liefern kann,
um die gefahr¨ lichsten Fehler bei Fahrmanov¨ ern im Kreuzungsbereich zu reduzieren. Im Vergleich
zu vorhandenen Warnungsassistenzsystemen kann ein solcher Ansatz die Akzeptanz des Nutzers
erhohen¨ und ist zudem kosteneffizient.
Durch die Identifikation der wichtigsten Einflussgroßen¨ auf die Fahraufgabe an einer Kreuzung und
durch die systematische Analyse ihre Auswirkungen auf Blickbewegungs-Strategien, Risiko Wahr-
nehmung und menschlicher Kognition wird eine Grundlage fur¨ die Entwicklung weiterer Assisten-
zsysteme auf der Fuhr¨ ungsebene gelegt.Abstract
This thesis focuses on relevant topics of traffic safety and future mobility by identifying intersections
as traffic situations with the highest potential for improvement of road safety. The increase of inner
city traffic, effects of aging societies, increasing mobility and reaching the limits of assistances in
the longitudinal traffic, brings intersections in the focus of future scientific research. Conventional
approachesinthedesignofIntersectionAssistanceslacktheuser’sacceptance,areoftenineffective
or present the high-cost solutions.
Within this thesis, an innovative, human-centered approach for error prevention in the traffic is given.
The analysis and modeling of driver cognition during the negotiation of an intersection is identified
as the crucial element for the development of an ergonomic Intersection Assistance. The performed
analysisinthecombinationwithadvancedresearchinthisareahasresultedinthefoundationforthe
computer simulation of driver cognition, which can be applied in the process of assistance develop-
ment. Experiments in the driving simulator environment, conducted within this work, identified the
substantial requirements for the support in intersection scenarios. Furthermore, these experiments
demonstratedthattheinformationintheapproachingsegmentofintersection,whichisbasedononly
navigationandon-boarddata,couldpreventthemostcriticalerrorsofintersectionperformance. The
useracceptancecan beincreasedincomparisontotheavailablewarningassistances, andpresents
a less costly solution than the automatic brake assistance.
Byidentifyingthemostimportantfactorsofthetaskperformanceatintersectionsandbyasystematic
analysis of their influence on applied visual strategies, risk perception, task demand and driver cog-
nitive processing, this thesis offers a groundwork for the future development of Driving Assistances
on the guidance level of the driving task.Glossary
ABS - Antiblock-locking System
ACC - Active Cruise Control or Adaptive Cruise Control
ADAS - Advanced Driver Assistance Systems
AOI - Area of Interest
C2C - Car to Car
C2I - Car to Infrastructure
C2X - Car to Car and Car to Infrastructure
CoP - Code of Practice
CWS - Collision Warning System
ESP - Electronic Stability Program
FoE - Focus of Expansion
FSM - Finite State Machine
GIDAS - German In-Depth Accident Study
HMI - Human Machine Interface
HUD - Head-up Display
ITS - Intelligent Transport Systems
IVIS - In-Vehicle Information Systems
LDW - Lane Departure Warning
LKA - Lane Keeping Assistance
LTM - Long-Term Memory
MAS - Multi-Agent System
SA - Situational Awareness
STM - Short-Term Memory
TCI - Task Capability Interface model from Fuller¨
TCS - Traction Control System
TTI - Time To Intersection
WM - Working MemoryContents
1 Introduction 1
1.1 Motivation and Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Approach, Contribution and Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 MethodologiesforDevelopmentofAdvancedDriverAssistanceSystems
forIntersections 6
2.1 Overview of Advanced Driver Assistance Systems . . . . . . . . . . . . . . . . . . . . 7
2.2 Ergonomic Approach for Development of Advanced Driver Assistance Systems . . . . 10
2.3 Intersection Assistance Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4 Defining Applications of Intersection Assistance . . . . . . . . . . . . . . . . . . . . . . 14
2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3 ClassificationofTrafficSituationsandDriverModelfortheTaskAnalysisofDriving 25
3.1 of Traffic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2 Relationship Between Driving Task and Driver Model . . . . . . . . . . . . . . . . . . . 32
3.3 Cognitive Representation for the Performance of Driving Task on the Guidance Level . 34
3.3.1 Information Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.3.2 Information Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Cognitive States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Cognitive Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3.3 Risk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.3.4 Task Demand and Workload . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4 EyeMovementsandVisualStrategiesContents 9
inDriving 63
4.1 Mechanisms of Eye Movements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.2 Visual Strategies in Driving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2.1 Visual Perception of Distance and Speed . . . . . . . . . . . . . . . . . . . . . 65
4.2.2 Visual Strategies at Intersection . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5 TheoreticalandExperimentalAnalysisofDrivingTaskatIntersections 74
5.1 Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.1.1 Driving simulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Driving Simulator Software: SILAB . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.1.2 Eye Tracking System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.2 The Objective and Selected Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.3 Theoretical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.3.1 Theoretical Task Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.3.2 Analysis of Visual Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.3.3 Task Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.4 Experimental Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.4.1 Test Sample and Experimental Procedure . . . . . . . . . . . . . . . . . . . . . 95
5.4.2 Presentation and Discussion of Results . . . . . . . . . . . . . . . . . . . . . . 97
Subjective data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Objective Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
5.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.5.1 Similarity and Difference in Applied Visual Strategies among Scenarios . . . . 129
5.5.2 Suggestion for Intersection Assistance System . . . . . . . . . . . . . . . . . . 132
5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
6 SurveyofExistingCognitiveDriverModels 138
6.1 Why Modeling the Driver? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.2 Driver Models in Cognitive Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.2.1 ACT-R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
ACT-R Driver Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.2.2 SOAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143DRIVER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
6.2.3 QN-MHP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 Driver Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
6.3 Stand Alone Cognitive Driver Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.3.1 COSMODRIVE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.3.2 ACME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.3.3 PELOPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.3.4 SSDRIVE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
6.3.5 Other models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
7 RecommendationfortheComputerSimulationofDriverCognition 155
7.1 Approach and Guidelines for the Model Development. . . . . . . . . . . . . . . . . . . 156
7.1.1 General Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
7.1.2 Modeling the Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
7.1.3 the States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
7.1.4 Modelling the Operative Cognitive Agents . . . . . . . . . . . . . . . . . . . . . 171
7.2 Implemented Functionality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
7.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
8 SummaryandFurtherWork 177
A ClassificationofIntersections 189
B Visualanalysis 198