Latent Variable Models of Attitudes and Preferences, and Their Prediction of Autonomous Vehicle Adoption Intent
Latent Variable Models of Attitudes and Preferences, and Their Prediction of Autonomous Vehicle Adoption Intent
Co-Principal Investigator: Ram M. Pendyala, Director, School of Sustainable Engineering and the Built Environment
Project Duration: 24 months
Project Budget (Federal UTC Funds): $59,703
Project Budget (Cost-share): $30,046
Abstract
In summer 2019, Dr. Sara Khoeini and Dr. Ram Pendyala launched a survey to understand how the market may perceive, adopt, and adapt to transformative transportation technologies mainly autonomous vehicles and mobility-on-demand services in the Phoenix Metropolitan area. The final survey was part of their study, Attitudes towards Emerging Mobility Options and Technologies – Phase 2: Pilot and Full Deployments. Briefly, the goal of their study was to examine attitudes and perceptions of emerging mobility options and technologies including autonomous vehicles and mobility-on-demand services that are bringing transformative changes in the transportation landscape. A major goal for the data collected from this survey was to measure and understand people’s attitudes towards and perceptions of these technologies and services in order to enhance transportation forecasting models.
The funded project proposal from Drs. Khoeini and Pendyala resulted in a dataset with N>1,000 with a series of targeted questions for following topics: (A) Attitudes and Preferences, (B) Household Vehicles and Residential Preferences, (C) Current Travel Patterns, (D) Mobility on Demand and Shared Mobility Services (Use and Attitudes), (E) Thoughts on Autonomous Vehicles (Familiarity, Perceptions, & Attitudes), and (F) Background Information (Demographics). This project proposal focuses on the analyses of this extensive dataset to examine (1) Dimensionality and Factor Structure of participant Attitudes and Preferences toward transportation and life in general (Section A) using both common factor models and latent class models, and (2) Prediction of aspects of participant thoughts on Autonomous Vehicles (Section E) with a particular focus on predicting attitudes regarding autonomous vehicle adoption (buying, leasing, expense).
Autonomous vehicles (AV) (also referred to as driverless cars or self-driving cars) are capable of navigating without human input using an array of technologies such as radar, lidar, GPS, odometry, and computer vision. Most industry experts suggest that autonomous vehicles will be on the road within a few years (Stoll, 2016). For example, the US Secretary of Transportation expects driverless cars to be in use all over the world by 2025 (Hauser, 2015), and The Institute of Electrical and Electronics Engineers (IEEE) predicts that up to 75% of all vehicles will be autonomous by 2040 (IEEE, 2012). In addition to the availability of AVs, ride-hailing companies, such as Uber and Lyft, have changed the transportation landscape as they provide door-to-door mobility-on-demand through the use of mobile apps.
Given these new transportation technologies and services, it is necessary for transportation forecasting models to account for market dynamics that will result from increased penetration of these technological innovations. Enhancing transportation forecasting models based on people’s attitudes toward and perceptions of these technologies and services is necessary. The data collected by Drs. Khoeini and Pendyala, provide the foundation for enhancing the forecasting models by capturing ample data on attitudes and perceptions, AV adoption expectations, ride-hailing use and attitudes, and background characteristics. Additionally, data from collaborating sites (Tampa, Atlanta, Austin) is available to examine the replicability of our models.
Research Products and Implementation
Final Report (coming soon)