Design, simulation, and implementation of a vision-based vehicle-following system

DSpace Repositorium (Manakin basiert)

Zur Kurzanzeige

dc.contributor.advisor Ruder, Hanns de_DE
dc.contributor.author Gehrig, Stefan de_DE
dc.date.accessioned 2000-09-04 de_DE
dc.date.accessioned 2014-03-18T10:08:26Z
dc.date.available 2000-09-04 de_DE
dc.date.available 2014-03-18T10:08:26Z
dc.date.issued 2000 de_DE
dc.identifier.other 086912046 de_DE
dc.identifier.uri http://nbn-resolving.de/urn:nbn:de:bsz:21-opus-1647 de_DE
dc.identifier.uri http://hdl.handle.net/10900/48114
dc.description.abstract Das Auto der Zukunft wird zunehmend intelligenter. Systeme wie das ABS sind heute bereits selbstverständlich. Der Bremsassistent, der von Daimler-Benz 1997 auf dem Markt eingeführt wurde, zeigt die Tendenz zu mehr Intelligenz im Auto an. Als weitere Stufe werden Assistenzsysteme für das Auto auf den Markt kommen, die den Fahrer teilweise von seinen Fahraufgaben entlasten. Solche Systeme arbeiten unter anderem mit Kamerasensorik. Die Dissertation beschäftigt sich mit dem Bildverstehen von Straßenszenen basierend auf Sensordaten eines Stereo-Kamerasystems. Die Hauptanwendung zielt auf ein autonomes Fahren im Stop-And-Go-Verkehr ab. Im Rahmen der Dissertation wird eine Situationsanalyse der Straßenszene für ein Fahrzeugfolge-Szenario vorgenommen, um Daten für die Längs- und Querführung des Fahrzeugs zu gewinnen. Die Arbeit ist in drei zentrale Punkte aufgeteilt. Um geeignet abstrahierte Daten zu gewinnen, wird an einigen Stellen neue Bildverarbeitungsalgorithmik entworfen. Um ein sicheres Fahrzeugfolgen zu garantieren wird zuerst ein Algorithmik entworfen, die ein exaktes Folgen des Pfades des vorausfahrenden Fahrzeigs ermöglicht. Zur temporalen Integration der gewonnenen Daten der Straß enszene wird eine Karte der Fahrzeugumgebung aufgebaut. Diese wiederum dient als Grundlage für eine Planungs- und Entscheidungsebene, die Steuervorgaben für die Fahrzeugführung (Ansteuerung von Bremse, Gas und Lenkung) generiert. Insbesondere wird auf dieser Ebene eine Hindernisvermeidungsstrategie implementiert, die es ermöglicht, dass allen erkannten Hindernissen ausgewichen wird. Dieses Verhalten wird durch Repräsentierung des Pfades zum vorausfahrenden Fahrzeug als elastisches Band erreicht. Die erkannten Hindernisse der Szene üben virtuelle Kräfte auf das Band aus und deformieren es, bis ein kollisionsfreier Weg erreicht ist. Falls das nicht möglich ist, wird solch eine Situation erkannt und eine manuelle übernahme seitens des Fahrers wird notwendig. de_DE
dc.description.abstract The motor vehicle of the future is equipped with an increasing number of intelligent systems. Anti-lock brakes (ABS) are considered standard today. The electronic stability program (ESP) system underlines the tendency towards more intelligence in the car. Further steps include driver assistance systems that facilitate driving by taking over parts of the driving task. For that purpose, the environment around the vehicle is perceived with appropriate sensors, such as cameras. This dissertation focuses on the image analysis of traffic scenes using a stereo camera system as the main sensor. Its main application is autonomous driving in stop-and-go traffic. This thesis considers situation assessment, planning, and decision. The resulting data are used by control algorithms to send commands to the actuators for longitudinal and lateral vehicle guidance. The thesis is divided into three parts. First, an algorithm is developed that allows the autonomous vehicle to exactly follow the path of a preceding vehicle. For this purpose, the paths of the autonomous vehicle and of the preceding vehicle are reconstructed in an inertial reference frame. Second, for the temporal integration of the acquired traffic scene data, a global map of the vehicle's environment is built. This map is the basis for the planning and decision stages that ultimately generate the data for longitudinal and lateral control. Third, the control commands which are issued to the actuators must guarantee a collision-free path. This collision avoidance capability is obtained using the notion of an elastic band for the path between the autonomous and the preceding vehicle. The elastic band is deformed by obstacles in its vicinity that exert forces on the band and consequently deform it. As a result, either a collision-free path is obtained or, in case of infeasible situations, a warning to the driver is issued who must take over the control of the vehicle manually. en
dc.language.iso de de_DE
dc.publisher Universität Tübingen de_DE
dc.rights ubt-podok de_DE
dc.rights.uri http://tobias-lib.uni-tuebingen.de/doku/lic_mit_pod.php?la=de de_DE
dc.rights.uri http://tobias-lib.uni-tuebingen.de/doku/lic_mit_pod.php?la=en en
dc.subject.classification Bildverarbeitung , Robotik de_DE
dc.subject.ddc 530 de_DE
dc.subject.other Intelligente Fahrzeuge , Pfadplanung de_DE
dc.subject.other Intelligent vehicles , Computervision , Robotics , Path planning en
dc.title Design, simulation, and implementation of a vision-based vehicle-following system en
dc.type PhDThesis de_DE
dc.date.updated 2004-10-29 de_DE
dcterms.dateAccepted 2000-07-31 de_DE
utue.publikation.fachbereich Sonstige - Mathematik und Physik de_DE
utue.publikation.fakultaet 7 Mathematisch-Naturwissenschaftliche Fakultät de_DE
dcterms.DCMIType Text de_DE
utue.publikation.typ doctoralThesis de_DE
utue.opus.id 164 de_DE
thesis.grantor 12/13 Fakultät für Mathematik und Physik de_DE

Dateien:

Das Dokument erscheint in:

Zur Kurzanzeige