Naturalistic Experiment to Simulate Travel Behavior Implications of Self-Driving Vehicles: The Chauffeur Experiment Mustapha Harb UC Berkeley UC Davis Georgia Tech October 24, 2018
Motivation: A future with fully selfdriving vehicles is fast approaching. Governments are racing to develop the necessary policies. How will fully autonomous vehicles impact travel and activity behavior? Todd Trumbull, The Chronicle 2
Predicting Travel Implications: Rich transport behavior literature: Pricing, multitasking, parking, sharing vehicles and rides, mode choice, safety, etc Approaches: Driving simulators and controlled testbeds Stated preference surveys Microsimulations and agent based model scenarios US DOE, Volpe center (August 19, 2014) The Apopka voice (July 26, 2016) ISI Foundation & ISI Global Science Foundation (April 20, 2016) 3
Research Findings on AV impacts: Drastically reduced vehicle fleet can serve current demand ( 10% of current fleet) (Fagnant & Kockelman, 2014; Fagnant et al., 2015; OECD, 2015) But vehicle miles traveled increases 8-10% vehicle relocation only (Fagnant & Kockelman 2014; Fagnant et al. 2015) 4-20% ease of driving, network efficiency (Gucwa, 2014; Childress et al. 2014) 5-35% depending on penetration and level of automation (Fehr&Peers, 2014) 6-90% depending on shared vehicles & rides, transit quality (OECD, 2015) People (today) are willing to pay for Automated Vehicles: $4,900 on average; ranges from $0 - $10,000+ (Daziano et al., 2016) There s hesitancy towards adoption and sharing self-driving car: 52% in US say they ll use an AV; 27% say they ll use a shared AV (WEF/BCG, 2015) 4
Naturalistic Experiment: FUTURE OF INTEREST: a fully autonomous vehicle Don t have to drive the car Full multitasking No parking worries Can send on errands SIMULATION OF FUTURE: a personal driver SCENARIO Status quo infrastructure and policies. Private vehicle ownership. Subjects use their own household vehicles. We provide the chauffeur to the household for free. Chauffeur stays with the car whether in use or not. 5
Naturalistic Experiment: Recruitment & Onboarding Travel Week 1: Status quo Travel Week 2: 60 hours of chauffeur service Travel Week 3: Status quo Offboarding Entrance Online survey All travel of subjects and vehicles tracked via mobile phone app, vehicle tracking GPS, and subject and chauffeur reports Exit Online survey 6
Summer 2017 Beta Test: Dense experiment run a beta test on 13 subjects before running the real experiment Five Retirees: Free time Safety concerns Travel restrictions (distance and time of day) Four Millennials: Tech savvy Uber/Lyft users Active lifestyle Four Families: Parents schedule is tied to kids schedule and vice versa 7
Average VMT by subject Finding 1: VMT Increased 600 500 400 300 21% 17 % 62 % 83 % 200 100 0 Non-chauffeur weeks (average) Chauffeur week Non-chauffeur weeks (average) Chauffeur week Non-chauffeur weeks (average) Chauffeur week Non-chauffeur weeks (average) Retirees Families Millennials Entire Sample Chauffeur week 8
Finding 2: Change in Activity Patterns 9
Beta Test Results Summary: (Harb et al., 2017) More auto travel 83% increase in VMT 21% of increased VMT were ghost trips Change in activity patterns 58% increase in # of vehicle trips 91% increase in # of longer trips (over 20 miles) 88% increase in # of evening trips (after 6 pm) Impact on miles walked was not directional 30% decreased (-31% on average), 70% increased (+38% on average) Virtually no biking, transit, TNC use in the sample 10
Potential Limitations: Resource limitations ($$$) Sample size Relatively short treatment/adjustment time (1 week) 60 hour chauffeur limit Novelty effect Context limitations Private ownership model, not shared Automation is provided for free Technology limitations Human driver, not robot 11
Quotes From Exit Interviews/Surveys: With all the limitations of the experiment, I definitely felt the benefits of a self-driving car. I noticed that I reach work less tired, I noticed that I can do work on my way back home and not worry much about traffic jams, and I noticed that my commute overall feels more pleasant. - Millennial I spend a lot of time in the car driving my kids around to activities. Having a self-driving car would enable me to spend more time on work and would afford my kids more freedom. - Family At my age, I am looking forward to the independence a self-driving car will provide as my driving skills decline. I believe self-driving cars will improve safety in driving, a real boon. - Retiree 12
Policy Implications: Self-driving cars will change how we travel. Different demographics will have unique changes in travel behavior. The technology will most likely increase overall VMT. The future cannot be stopped, but it can be improved! Need for policies to maximize the societal benefits and minimize the negative externalities More studies to gain further insight on the implications of the technology on travel behavior and the transport system. 13