Project: #364 Autonomous Food Delivery Impacts on Traffic and Sustainability Progress Report - Reporting Period Ending: Sept. 30, 2021 Principal Investigator: Destenie Nock Status: Active Start Date: July 1, 2021 End Date: June 30, 2022 Research Type: Applied Grant Type: Research Grant Program: FAST Act - Mobility National (2016 - 2022) Grant Cycle: 2021 Mobility UTC Progress Report (Last Updated: Sept. 29, 2021, 11 a.m.) % Project Completed to Date: 30 % Grant Award Expended: 0 % Match Expended & Document: 0 USDOT Requirements Accomplishments Major goals of the project include the following: • Produce a traffic modelling tool for understanding how grocery delivery market penetration will impact congestion and environmental sustainability • Environmental sustainability and equity trade-off assessment. We will perform an analysis of how grocery delivery impacts air pollution and congestion in low income areas. • Develop a relationship with grocery and autonomous vehicle companies to promote interdisciplinary transportation planning. Over the past few months we have improved the grocery demand model and calibrated it against real traffic counts. Our demand model simulates grocery delivery in a built environment. It estimates number of trips and vehicle miles traveled (VMT) under different grocery batching scenarios and market penetration rates. The batching scenarios refer to a grocery store combining different household deliveries into a single grouping. The market penetration rate is the proportion of grocery shoppers that switch from in-store shopping to online delivery orders. Over the past few months several improvements, analysis, and changes have been made to improve the demand model. These include the following: 1. We improved the model by only performing delivery trips based on the distance the grocery store is to the home. This is the closest grocery to home system which assumes that deliveries are being made to the closest traffic analysis zone (TAZ) which has a grocery store. 2. We have been able to better simulate grocery demand because of new tour-level information. A tour refers to the chain of trips that a person makes. With this level of information, we can determine where people’s homes are, and can route delivery tours to actual demand locations. 3. We can generate representative demand scenarios that serve as inputs to commercial transportation analysis software. The outputs from the demand model are in the form of Origin-Destination (OD) matrices, that result from a specified batch size and market penetration rate. In addition to developing the demand model, we have fitted it into a Monte Carlo simulation to observe trends and uncertainty in our delivery system after several iterations. With this, we can choose OD matrices near the median and the extreme, and them use them as input to transportation analysis software. This will allow us to have a bound on congestion impacts for a certain batch size and penetration rate. Ongoing work: Ultimately the results of the demand model will be fed into the network model, which will allow us to estimate the impact that grocery delivery has on congestion and the spatial distribution of GHG emissions associated with the transportation sector. The network model of choice is transportation analysis software, Visum. Some of the benefits of Visum include tracking trip speed, count, and destination. Visum takes TAZ trip tables as an input and based off the trip counts it returns an output. We have obtained an academic license for Visum, and are currently calibrating the model to reflect current levels of congestion in the Seattle metropolitan area. Once the calibration is complete we will test the impact of different market penetrations and batch scenarios to see how they might impact the transportation system at large. Figure 1. Median and extremes of the distribution of the impact of grocery delivery on VMT for different batch sizes (i.e. # of deliveries per trip). Collaborations: As a part of this project we have met with representatives from Giant Eagle (grocery store), EasyMile (autonomous food delivery company), and the Pugent Sound Regional Council (transportation planner). Impacts The effectiveness of the transportation system will be improved by understanding how food delivery impacts congestion, and air pollution emissions. Other We have accomplished the following during the reporting period: • We developed food delivery demand model(s) and use systems transportation modeling approaches to quantify how autonomous food delivery could affect peak hour traffic operations, and determine delivery alternatives that are most sustainable • We are developing relationships with relevant stakeholders including Giant Eagle (local grocery store), Puget Sound Regional Council (transportation authority), and autonomous vehicle companies Outcomes New Partners Puget Sound Regional Council (transportation authority) Issues - During the Spring 2021 Semester our first PhD student Anthony Reid decided to leave the program with an MS degree due to a combination of personal reasons and the COVID pandemic. This has been resolved by hiring a second PhD student (Mateo) who is now leading the analysis on the project.