Research Projects

Stop the Video

Research Projects

STATUS: Complete YEAR: 2021 TOPIC AREA: Connected and autonomous systems E-commerce and urban freight CENTER: PSR

Evaluation of Sidewalk Delivery Robot Interactions with Pedestrians and Bicyclists

Project Summary

Project number: PSR-21-16
Funding source: USDOT
Contract number: 69A3551747109
Total cost: $97,421
Performance period: 8/15/2021 to 8/14/2022

Project description

The commercial deployment of autonomous sidewalk delivery robots throughout the campus of Northern Arizona University (NAU) offers students, staff, and faculty a convenient option for food delivery to their dormitory or office doorsteps. However, these delivery robots, which were first deployed at NAU in March 2019 and have continued providing critical services during the ongoing COVID-19 pandemic, travel across campus along sidewalks and pathways already utilized by pedestrians and bicyclists, creating a potential for conflicts amongst the different pathway users and potentially unsafe conditions. This project seeks to assess the interactions between this autonomous delivery vehicle fleet and active travelers by collecting and analyzing passive video and intercept survey data. Specifically, this study will adapt a surrogate safety measure to identify the prevalence of observed conflicts between autonomous sidewalk delivery robots and human pathway users and also estimate the predictors of this objective outcome in addition to the perceived safety risk of interactions measured with video-embedded survey questioning.

To address the stated research objectives, this study will analyze observed and self-reported data collected from two primary sources. First, video cameras will be installed at five locations where the robot delivery fleet steadily interacts with high volumes of pedestrians and bicyclists to document interactions and conflicts. Video collection will occur daily at morning, midday, and evening mealtime periods over two weeks in Fall 2021. These data will be reviewed and coded by two independent researchers to adapt a surrogate safety measure of post encroachment time in calculating episodes of conflict severity. Additionally, since not all interactions will result in a quantitative measure of conflict, those instances of interaction that result in evasive actions such as an abrupt change in speed or travel direction by pathway users will be documented. Statistical modeling will be conducted to identify the design, temporal, and environmental determinants of pedestrian and bicyclist safety-in terms of conflict frequency or severity-when interacting with robot delivery vehicles on pathways. Second, a tablet-based intercept survey instrument will be designed and administered by Spring 2022 at high-activity locations on the NAU campus. This original survey instrument will record NAU faculty, staff, and student responses to questions about their socioeconomic background, travel options and patterns, views of and experiences with NAU sidewalk delivery robots, and perceptions of security and comfort when watching short video clips of prerecorded interactions with a sidewalk delivery robot from the perspective of a pedestrian or bicyclist. Responses to Likert-scale survey questions on perceived collision risk for active travelers will be assessed by estimating ordinal logistic regression models for each human pathway user type as a function of additional characteristics measured from survey questions.

The implementation of this proposed study design will provide novel insights on how the observed and perceived active travel behaviors of a university population have been impacted by the introduction of smart autonomous freight vehicles sharing their sidewalks and bike paths. As emerging transportation technologies continue to vie for the use of public spaces and facilities, studies investigating the safety, security, and comfort related to increased conflicts and interactions with the most vulnerable users of our transportation systems will become increasingly important to transportation engineers and planners as they redesign facilities to safely accommodate the higher demand for their use. This proposed study will produce early evidence to help shape future mitigation strategies and transportation policies for safely operating automated sidewalk delivery robots in public settings.


Steven Gehrke
Assistant Professor
Room 252 Building 70 SBS West
Flagstaff, AZ 86011
United States
[email protected]