from the University of Texas at Austin and “What problems might arise?“This stupidity on the part of the department reminds me of how smoothly the evacuation of New Orleans went during Hurricane Katrina in 2005. I was surprised that anyone spent time and money on this. Thanks to Charles for the diagram that depicts my understanding of the process view. ——Anthony.
Escape a hurricane in a driverless car
Supercomputer-aided simulation to aid car-free crowds in Houston evacuation scenario
When a hurricane strikes, the most vulnerable people are not always able to evacuate in time. Scientists at the University of Texas at Austin are using supercomputers at the Texas Advanced Computing Center (TACC) to study how shared autonomous vehicles (SAVs) can allow people without their own cars to enter shelters and avoid harm.
“One of the key findings from our study is that if you need to size the system during an evacuation, you're going to need to scale the long stretch of coastline between Galveston and Houston,” said Kara Kockelman. There is one shared vehicle for every 14 evacuees, she co-authored the study, published in the journal. Transportation Planning and Technology (June 2024).
The idea is to use shared self-driving vehicles, similar to the robotaxis from companies like Cruise and Waymo, to bring carless people to bus stops and then transport them to hurricane shelters in Houston's interior.
“In these low-density, rural settings, it can be difficult to complete this task in less than a few hours,” Cockelman added. High-risk areas for evacuation include Brazoria County, Chambers County, Galveston, Harris and Liberty counties. The study focuses on the thousands of people who may be stranded, such as those listed in Medicare databases who don't have cars or access to transportation.
When a Category 5 hurricane strikes, 900,000 people are expected to receive evacuation orders, accounting for approximately 12.4% of the total population of the Houston area. Engineers estimate that background traffic is about 50% of normal traffic load. Assume the rest of the population remains where it is.
Evacuees will traverse Houston's complex road network; 36,124 links distributed in 5,217 areas (called traffic analysis areas). Of these, 1,035 areas are at high risk of being hit by a severe hurricane.
Kockelman received the assignment for his work on traffic simulations using SAV on TACC's Frontera supercomputer, the fastest supercomputer on a U.S. college campus and funded by the National Science Foundation.
“Without supercomputers, this work would not be possible,” Kokelman added. “We track individuals and vehicles every few seconds in real traffic over 24 hours or days, as traffic changes from morning to evening and through the night.”
Transportation decisions in the simulation take traffic conditions into account, choosing the best route and prioritizing pickups to minimize the time needed to evacuate people.
Engineers use (SUMO) Simulation of Urban Mobility traffic simulation software to assess traffic congestion and network capacity. They simulated pre-disaster evacuation scenarios in the days before the hurricane made landfall.
“SUMO simulates the daily activities of everyone in the area,” said Kentaro Mori, a UT Austin doctoral student supervised by Kockelman. “There’s a lot of complexity that adds to the computational cost. Without TACC, we wouldn’t be able to run many of the scenarios we need to really answer these important research questions and make the best policy recommendations.
The team started the simulation with 200 robotaxis and later expanded to 1,200 – and they also tried different sizes of cars.
“Ultimately, the five-passenger car is the most flexible,” Cokerman said. “These vehicles accelerate into traffic faster than larger vehicles.” What's more, simulations show diminishing returns for a fleet of more than 200 ships.
This study is groundbreaking work that provides data to recommend viable alternative modes of transportation for hurricane evacuees who do not have access to private vehicles. While the results have not been directly used in hurricane applications, engineers did consult with Texas evacuation leaders in forming the study. The author predicts that as companies such as Waymo expand their passenger capacity, SAVs will play a greater role in evacuation; operations will be improved through intelligent repositioning of idle vehicles, optimal dynamic ridesharing matching, and enhanced pathfinding algorithms. .
Kokelman added that the traffic simulation could be adapted to other cities and different disaster evacuation scenarios, such as wildfires on the West Coast.
“The ability to simulate in detail and take into account the uncertainty and heterogeneity in a population, in terms of how people make decisions or how traffic flows unfold on a second-to-second basis, was not previously possible,” Cokerman said. That richness comes to life through the use of the TACC system. We're lucky that we're at UT Austin and can take advantage of this capability.
The study, “Utilizing shared autonomous vehicles for vulnerable populations during pre-disaster evacuations,” was published in the journal Transportation Planning and Technology in June 2024. The study authors are Jooyong Lee, Department of Urban and Transportation Engineering, Kyunggi University, Suwon, South Korea; and Kara M. Kockelman, Department of Transportation Engineering, Department of Civil, Architectural and Environmental Engineering, University of Texas at Austin.
Relevant
