عنوان

Driver Behavior Modeling for Autonomous Vehicle Motion Planning

Number

TLpq2385624822

.Language of Text, Soundtrack etc

انگلیسی

Title Proper

Driver Behavior Modeling for Autonomous Vehicle Motion Planning

General Material Designation

[Thesis]

First Statement of Responsibility

Ramyar, Saina

Subsequent Statement of Responsibility

Homaifar, Abdollah

Name of Publisher, Distributor, etc.

North Carolina Agricultural and Technical State University

Date of Publication, Distribution, etc.

2019

Specific Material Designation and Extent of Item

108

Dissertation or thesis details and type of degree

Ph.D.

Body granting the degree

North Carolina Agricultural and Technical State University

Text preceding or following the note

2019

Text of Note

The advanced driving assistance systems (ADAS) such as lane departure warning and collision avoidance have great potentials in improving traffic safety. In order for the ADAS to be able to detect and prevent an accident, it is required to predict other vehicles' actions and plan the subject vehicle's motion accordingly. Due to the complexity of human-vehicle interaction, obtaining an explicit model for analyzing the drivers' behaviors is difficult. Instead, models are developed for various driver decisions and driving scenarios which are integrated together. In this dissertation, machine learning models are developed to represent human driver behaviors both on urban roads and highways. A fuzzy clustering approach is proposed to predict the driver's actions at intersections. Moreover, an anomaly detection technique is used to identify potential abnormalities that may lead to various hazards during the process of a lane change. In addition to driver models developed for safe trajectory generation, a personalized driver model is proposed, which is learned through demonstration and performs according to the driver's preference. Furthermore, a cooperative car following model for platoons is proposed which estimates the preceding vehicle's acceleration in case of communication failure and enables the platoon to maintain a relatively small inter-vehicle gap and remain string stable. All the models in this dissertation are based on naturalistic driving behavior and have been tested in various scenarios. The simulation results show the high accuracy of the proposed models and validate their applicability for autonomous motion planning.

Artificial intelligence

Robotics

Transportation

Homaifar, Abdollah

Ramyar, Saina

Electronic name

p

[Thesis]

276903

a

Y