On the Influence of Action Preparation on Steering Performance in a Lane Change Task
Loading...
Date
2011-01-12
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Kognitiv psychologische experimentelle Untersuchungen der Handlungsvorbereitung
und der Bewegungsausführung finden gemeinhin unter hoch kontrollierten Bedingungen im
Labor statt. Es ist allgemein anerkannt, dass adäquate zeitliche und inhaltliche
Handlungsvorbereitung die Reaktion des Individuums unterstützen, was sich z.B. in
verkürzten Reaktionszeiten auf imperative Stimuli zeigt. Verkehrspsychologische Forschung
wird dagegen häufig unter realitätsnahen Bedingungen, z. B. im Feld bzw. in modernen
Fahrsimulatoren durchgeführt. In diesem Forschungsbereich ist kaum untersucht, inwieweit
Mechanismen der Handlungsvorbereitung die Reaktionen eines Fahrers unterstützen. Trotz
der unterschiedlichen Ansätze erscheint das Konzept der Handlungsvorbereitung als durchaus
anwendbar auf fahrerische Lenkbewegungen, da das Fahren an sich die stete Antizipation
direkt bevorstehender bzw. zukünftiger Verkehrssituationen erfordert. Ausgehend von dieser
Annahme verfolgt diese Dissertation drei Ziele. Erstens galt es, grundlagenorientierte
experimentelle Forschungsparadigmen der Handlungsvorbereitung, insbesondere zeitliche
und ereignisbezogene Vorbereitung, in eine anwendungsorientierte Fahraufgabe zu
integrieren. Zweitens sollten grundlagenorientierte kognitionspsychologische Vorhersagen in
dieser angewandten Umgebung getestet werden, um entsprechende Befunde zu replizieren
bzw. zu erweitern. A priori besteht keine Sicherheit darüber, ob zeitliche und
ereignisbezogene Vorbereitung tatsächlich zu mit den Grundlagen vergleichbaren
Ergebnissen in einer Fahraufgabe führen, da sich eine kontinuierliche Fahraufgabe mit
komplexen beidhändigen Lenkbewegungen deutlich von einer Einzelfingerreaktion in einer
typischen Laboraufgabe unterscheidet. Drittens sollte eines der grundlagenorientierten
Paradigmen modifiziert und anhand einer angewandten Fragestellung eingesetzt werden, um
die Nützlichkeit grundlagenorientierter Forschungsparadigmen zur Untersuchung konkreter
Probleme beispielhaft zu demonstrieren.
Zu diesem Zweck wurde eine Spurwechselaufgabe in einem Labor für virtuelle Realität
entwickelt. Mit Hilfe dieser experimentellen Fahraufgabe wurden zeitliche und
ereignisbezogene Handlungsvorbereitungsprozesse in einer ersten Serie von fünf
Experimenten untersucht. Zeitliche Vorbereitung wurde im Rahmen eines variablen
Vorperiodenparadigmas implementiert, ereignisbezogene Vorbereitung in Übereinstimmung
mit dem Movement Precuing bzw. dem Response Priming Paradigma. Neben der
Reaktionszeit auf imperative Spurwechselsignale wurden kinematische Eigenschaften der Spurwechsellenkbewegung analysiert. Dazu wurde die gesamte Spurwechsellenkbewegung
in drei Einzelbewegungen unterteilt, von denen letztlich die ersten beiden anhand ihrer
Geschwindigkeitsprofile ausgewertet wurden. Grundsätzlich wurde vorhergesagt, dass
zeitliche und ereignisbezogene Handlungsvorbereitung Reaktionszeitverkürzungen zur Folge
haben und somit zu der Grundlagenliteratur ähnlichen Reaktionszeitmustern führen sollten.
Darüber hinaus und übereinstimmend mit der "movement integration"-Hypothese (Adam et
al., 2000) bzw. Ergebnissen von van Donkelaar und Franks (1991a) wurden kinematische
Effekte um das erste Lenkwinkelmaximum der Spurwechsellenkbewegung erwartet.
Schließlich dienten ein sechstes und ein siebtes Experiment der Untersuchung des
angewandten Problems, ob ein 3-dimensionales kontaktanaloges Head-Up Display (HUD)
die Handlungsvorbereitung effektiver als ein 2- oder 2½-dimensionales HUD unterstützt.
Die Ergebnisse der Experimentalserie waren mannigfaltig. Im Hinblick auf die ersten
fünf Experimente wurden grundsätzlich die aus den Grundlagen bekannten
Reaktionszeitergebnisse repliziert. Spezifischer betrachtet zeigten sich deutlichere Effekte
ereignisbezogener als zeitlicher Handlungsvorbereitung. Das Movement Precuing führte zum
bekannten "precuing effect", d.h. mit zunehmender Vorinformation über einen anstehenden
Spurwechsel verkürzten sich die Reaktionszeiten. Das Response Priming zog den bekannten
"validity effect" nach sich, d.h. die Reaktionszeiten fielen für neutral vorbereitete
Spurwechsel länger als für valide und kürzer als für invalide vorbereitete Spurwechsel aus.
Zeitliche Vorbereitung führte zwar immer zu kürzeren Reaktionszeiten als gar keine
Vorbereitung, aber die Unterschiede zwischen verschiedenen Vorperioden fielen im
Gegensatz zu Grundlagenerkenntnissen nur gering aus. In zwei Experimenten ergaben sich
zudem instabile Interaktionen zwischen zeitlicher und ereignisbezogener
Handlungsvorbereitung. Diese Ergebnisse deuten daraufhin, dass sich beide Arten der
Vorbereitung gegenseitig beeinflussen können. Im Gegensatz zu zeitlicher
Handlungsvorbereitung hatte ereignisbezogene Handlungsvorbereitung einen systematischen
Einfluss auf kinematische Eigenschaften der Lenkbewegung. Movement Precuing und
Response Priming neigten zuverlässig dazu, die Länge der beiden kinematischen Phasen um
das erste Lenkwinkelmaximum zu verkürzen. Diese Verkürzung wurde als erhöhte Effizienz
der Lenkbewegung interpretiert, da sich keine zusätzlichen Kosten im Sinne einer
schlechteren Steuerqualität ergaben. Die Ergebnisse der Experimente sechs und sieben fielen
nicht hypothesenkonform aus. Anstelle des 3-dimensionalen HUDs unterstütze das 2-dimensionale HUD die Handlungsvorbereitung am effektivsten, was unter Umständen den
überlernten einfachen zweidimensionalen Pfeilsymbolen zugeschrieben werden kann.
Zusammenfassend belegt diese Dissertation die Relevanz der
Handlungsvorbereitungskonzepte für Fahrmanöver. Die Reaktionszeitergebnisse weisen auf
vergleichbare Mechanismen der Informationsverarbeitung bei einfachen
Reaktionszeitaufgaben und komplexen, kontinuierlichen bimanuellen Steueraufgaben hin.
Darüber hinausgehend scheint ereignisbezogene Handlungsvorbereitung auch der
Optimierung der Bewegungsausführung zu dienen. Basierend auf diesen Hinweisen könnten
sich weiterführende Studien auf die Interaktion zwischen zeitlicher und ereignisbezogener
Handlungsvorbereitung, auf zusätzliche Fahrmanöver in realistischeren Szenarios oder auf
den Nutzen dieser Spurwechselaufgabe als diagnostisches Instrument konzentrieren.
Cognitive psychological experimental work on response preparation is generally conducted under highly controlled laboratory conditions. Adequate temporal as well as eventspecific preparation typically support the individual’s response preparation, for example in terms of reduced reaction times (RT) on imperative stimuli. By contrast, psychological research on driving tends to be conducted under realistic conditions, for example in the field or in sophisticated driving simulators. This research has largely overlooked the extent to which mechanisms of action preparation support the driver's reactions. Despite these two different approaches, the concept of action preparation should be highly applicable to vehicle steering, since driving requires the driver continuously to anticipate immediate and future traffic situations. Starting from this assumption, this dissertation’s first aim is to transfer basic experimental paradigms on action preparation – that is, temporal and event-specific preparation – into a driving task. Second, the predictions derived from cognitive psychological groundwork will be tested in this applied setting in order to replicate and enhance the corresponding findings. A priori it was not clear whether temporal and eventspecific preparation would lead to comparable results in a driving task, since a continuous driving task with complex bimanual steering movements is quite different from discrete single finger reactions in a typical laboratory task. Third, one of the basic paradigms on action preparation will be modified and transferred to an applied question in order to exemplify its usefulness for concrete problems in driving research. For these purposes, a lane change task was developed in a virtual reality environment. With the help of this experimental driving task, processes of temporal preparation as well as of event-specific preparation were examined in a first series of five experiments. Temporal preparation was implemented according to a variable foreperiod paradigm, while eventspecific preparation was realized according to the movement precuing and the response priming paradigm. Aside from measuring the reaction time (RT) on the imperative lane change signals, kinematic properties of the lane change steering wheel movement were evaluated. To this end, the entire lane change movement was divided into three submovements, the first two of which were eventually analyzed in terms of their velocity profiles. It was generally predicted that temporal and event-specific response preparation would reduce RTs and lead to result patterns comparable with evidence in the respective cognitive psychological literature. In keeping with the "movement integration"-hypothesis (Adam et al., 2000) and in accordance with the findings of van Donkelaar and Franks (1991a), kinematic effects were expected around the first peak steering wheel angle of the lane change steering wheel movement. Eventually, the sixth and seventh experiment served to examine the applied question whether a 3-dimensional conformal head-up-display (HUD) supports response preparation more effectively than a 2- or 2½-dimensional HUD. This experimental series produced a variety of results. With regard to the first five experiments, RT patterns known from basic research were replicated. More specifically, the effects of event-specific preparation were more pronounced than those of temporal preparation. Movement precuing led to the “precuing effect” – i.e., the more advance information on a lane change was available, the shorter the RT. Response priming led to the "validity effect": RT on neutrally prepared lane changes was longer than RT on validly prepared lane changes and shorter than RT on invalidly prepared lane changes. Temporal preparation always led to shorter RT than no preparation at all. But, contrary to the literature, the differences between different foreperiods were only shallow. Although unstable, interactions between temporal and event-specific preparatory processes occurred in two experiments. These results support the idea that both processes might mutually influence each other. Temporal preparation did not influence the kinematic properties of the steering wheel movement systematically, although event-specific preparation did. Movement precuing and response priming tended reliably to shorten the duration of the two kinematic phases centered around the first peak steering wheel angle. Since this shortening was not accompanied by increased costs in terms of lower steering quality, this modification was interpreted as increased movement efficiency. The results of Experiment 6 and 7 were not compliant with the hypotheses. Instead of conformal 3-dimensional HUDs, 2-dimensional HUDs tended to support response preparation most effectively. This effect is possibly due to the overlearned nature of simple 2-dimensional arrows. In conclusion, the thesis offers evidence for the relevance of concepts of response preparation for driving maneuvers. The RT results point to comparable mechanisms of information processing in simple RT tasks and more complex continuous bimanual steering tasks. Event-specific response preparation also seems to optimize response execution. Based on this evidence, future directions of this research might consist in further examining interactions between temporal and event-specific preparation, analyzing additional driving maneuvers in more realistic scenarios or making use of the lane change task as potential diagnostic tool.
Cognitive psychological experimental work on response preparation is generally conducted under highly controlled laboratory conditions. Adequate temporal as well as eventspecific preparation typically support the individual’s response preparation, for example in terms of reduced reaction times (RT) on imperative stimuli. By contrast, psychological research on driving tends to be conducted under realistic conditions, for example in the field or in sophisticated driving simulators. This research has largely overlooked the extent to which mechanisms of action preparation support the driver's reactions. Despite these two different approaches, the concept of action preparation should be highly applicable to vehicle steering, since driving requires the driver continuously to anticipate immediate and future traffic situations. Starting from this assumption, this dissertation’s first aim is to transfer basic experimental paradigms on action preparation – that is, temporal and event-specific preparation – into a driving task. Second, the predictions derived from cognitive psychological groundwork will be tested in this applied setting in order to replicate and enhance the corresponding findings. A priori it was not clear whether temporal and eventspecific preparation would lead to comparable results in a driving task, since a continuous driving task with complex bimanual steering movements is quite different from discrete single finger reactions in a typical laboratory task. Third, one of the basic paradigms on action preparation will be modified and transferred to an applied question in order to exemplify its usefulness for concrete problems in driving research. For these purposes, a lane change task was developed in a virtual reality environment. With the help of this experimental driving task, processes of temporal preparation as well as of event-specific preparation were examined in a first series of five experiments. Temporal preparation was implemented according to a variable foreperiod paradigm, while eventspecific preparation was realized according to the movement precuing and the response priming paradigm. Aside from measuring the reaction time (RT) on the imperative lane change signals, kinematic properties of the lane change steering wheel movement were evaluated. To this end, the entire lane change movement was divided into three submovements, the first two of which were eventually analyzed in terms of their velocity profiles. It was generally predicted that temporal and event-specific response preparation would reduce RTs and lead to result patterns comparable with evidence in the respective cognitive psychological literature. In keeping with the "movement integration"-hypothesis (Adam et al., 2000) and in accordance with the findings of van Donkelaar and Franks (1991a), kinematic effects were expected around the first peak steering wheel angle of the lane change steering wheel movement. Eventually, the sixth and seventh experiment served to examine the applied question whether a 3-dimensional conformal head-up-display (HUD) supports response preparation more effectively than a 2- or 2½-dimensional HUD. This experimental series produced a variety of results. With regard to the first five experiments, RT patterns known from basic research were replicated. More specifically, the effects of event-specific preparation were more pronounced than those of temporal preparation. Movement precuing led to the “precuing effect” – i.e., the more advance information on a lane change was available, the shorter the RT. Response priming led to the "validity effect": RT on neutrally prepared lane changes was longer than RT on validly prepared lane changes and shorter than RT on invalidly prepared lane changes. Temporal preparation always led to shorter RT than no preparation at all. But, contrary to the literature, the differences between different foreperiods were only shallow. Although unstable, interactions between temporal and event-specific preparatory processes occurred in two experiments. These results support the idea that both processes might mutually influence each other. Temporal preparation did not influence the kinematic properties of the steering wheel movement systematically, although event-specific preparation did. Movement precuing and response priming tended reliably to shorten the duration of the two kinematic phases centered around the first peak steering wheel angle. Since this shortening was not accompanied by increased costs in terms of lower steering quality, this modification was interpreted as increased movement efficiency. The results of Experiment 6 and 7 were not compliant with the hypotheses. Instead of conformal 3-dimensional HUDs, 2-dimensional HUDs tended to support response preparation most effectively. This effect is possibly due to the overlearned nature of simple 2-dimensional arrows. In conclusion, the thesis offers evidence for the relevance of concepts of response preparation for driving maneuvers. The RT results point to comparable mechanisms of information processing in simple RT tasks and more complex continuous bimanual steering tasks. Event-specific response preparation also seems to optimize response execution. Based on this evidence, future directions of this research might consist in further examining interactions between temporal and event-specific preparation, analyzing additional driving maneuvers in more realistic scenarios or making use of the lane change task as potential diagnostic tool.
Description
Table of contents
Keywords
Response preparation, Movement precueing technique, Steering dynamics, Lane change task