Document Type

Thesis

Date of Degree

Spring 2010

Degree Name

MS (Master of Science)

Degree In

Industrial Engineering

First Advisor

Linda N. Boyle

Abstract

Trust in automatic controllers may have an impact on awareness of system limitations and ultimately use of these systems. The purpose of this research is to determine the effects and likelihood of too much trust (overtrust) in drivers that use adaptive cruise control (ACC), a type of automatic controller that maintains vehicle speed and headway time. To add to the existing literature, this study aimed to define a significant relationship among trust, use, and awareness of ACC limitations. A post mailed or electronic-based survey was distributed to potential ACC users with 118 responses used for the main analysis. The survey responses provided demographic information, illustrated levels of trust, awareness of ACC limitations, and system use. A hierarchical cluster analysis of the data related to trust in ACC produced four clusters: overtrust, cautious, neutral, and distrust. Binary and multinomial logistic regression models then predicted the likelihood for overtrust (cluster membership). Participants in the overtrust cluster displayed the lowest level of awareness regarding ACC limitations and the highest levels of misuse. Users were more likely to overtrust ACC if they were male, misused the system, lacked limitation awareness, and indicated a willingness to opt for ACC in their next vehicle. Overtrust in ACC was shown to lead to misuse and cloud awareness of the systems' limitations. Trust seemed to be formed in the initial interactions with ACC therefore a priori trust levels of potential users should be sought before appropriate guidance is given or demonstrated through test use.

Keywords

Adaptive Cruise Control, Automation, Cluster Analysis, Overtrust, Questionnaire, Trust

Pages

vi, 52 pages

Bibliography

Includes bibliographical references (pages 48-52).

Copyright

Copyright 2010 David Alexander Dickie

Share

COinS