Naturalistic Experiment in California Found Fully Automated Vehicles Increased Households' Vehicle Miles Traveled by 60 Percent, but that Overall Shifts Towards Car-Dependent Travel are Undesirable.

A Study Utilizing a Chauffeur Service to Mimic Fully Automated Vehicle Travel for Selected Households in Sacramento, CA Saw an Increase in Overall Vehicle Miles Traveled, Over Half of Which Being Due to Zero Occupancy Vehicle Trips.

Date Posted

Simulating Life with Personally-Owned Autonomous Vehicles through a Naturalistic Experiment with Personal Drivers

Summary Information

While fully automated vehicles (FAVs) offer convenience in performing daily tasks and running errands, they may have negative impacts on public transit, congestion, and pollution as they heavily promote car travel. In this study, a small sample of households in the Sacramento region, CA representing diverse demographics, modal preferences, mobility barriers, and weekly Vehicle Miles Traveled (VMT) were provided personal chauffeurs for one or two weeks to simulate travel behavior with a personally-owned FAV per household, in a period from August 2019 to March 2020, including the non FAV control weeks before and after the experiment.


In this study, 43 households were selected from a sampling frame of 4,010 households that had previously participated in a travel survey and agreed to be contacted for follow-up studies. Using a smartphone app and a GPS device, participants recorded their travel diaries during a two-week naturalistic experiment with a chauffeur mimicking the FAV. During the first (control) week prior to the experiment, travel diaries were recorded under no FAV conditions. Then, households received one or two weeks of the chauffeur service. In total, 34 households received 60-hours of chauffeur service for one week and nine households received 60 hours per week for two weeks, with the goal of understanding if changes in travel behavior persist as the treatment period is extended. After the chauffeur week(s), travel diaries were recorded for a second control week. The experiment identified trips with the chauffeur as Zero Occupancy Vehicle (ZOV) trips if the only individual in the car was the chauffeur.


  • The results indicated that the total number of trips increasing on average by 25 percent, 85 percent of which were ZOV trips, during the chauffeur week(s). The average VMT for all household vehicle trips as well as trips using non-household vehicles (e.g., Uber, car from work, friend’s car) was increased by over 60 percent, over half of which came from ZOV trips. The findings revealed a potential shift towards car-dependent travel for the participating households and a decrease in the use of other travel modes. Transit, ridehailing, biking, and walking trips dropped by 70 percent, 55 percent, 38 percent, and 10 percent, respectively.
  • Access to a FAV provided benefits to households with seniors and people with disabilities; households with retirees experienced the largest increase in VMT by 121 percent. Through the use of FAVs, elderly participants and those with disabilities reported substantial lifestyle improvements, including improved flexibility to travel at night, take longer distance trips and travel without being tied to a fixed transit schedule.
  • The households belonging to the lowest VMT category, dominated by single-person and elderly households, experienced the highest increase in VMT by 137 percent, indicating the advantage of having a FAV facilitating a more active lifestyle.
  • Despite an overall increase of 114 percent in household VMT for “AV” (chauffeur car) use, the VMT in traditional, non-automated vehicles dropped by 53 percent, as a single FAV was able to serve the entire household's travel needs. This finding suggests that FAV ownership has the potential to reduce the number of vehicles owned by households.

Simulating Life with Personally-Owned Autonomous Vehicles through a Naturalistic Experiment with Personal Drivers

Simulating Life with Personally-Owned Autonomous Vehicles through a Naturalistic Experiment with Personal Drivers
Source Publication Date
Harb, Mustapha; Jai Malik; Giovanni Circella; and Joan L. Walker
Prepared by Institute of Transportation Studies, Berkeley and Davis for the University of California Institute of Transportation Studies
Other Reference Number
Results Type
Deployment Locations