Abstract
Travel behavior in route choices under incidents will be modeled based on a disutility function for individuals and the calibrated regional network model. We will design a methodology to simulate the traffic and estimate dynamic O-D demand on the real time basis. The simulation adopts the historical traffic demand as an initial (base) demand and their pre-scribed route choices from the dynamic network model in the existing meso-scopic simulation tool (MAC-POSTS).
Description
Objective
Non-recurrent traffic congestion caused by roadway construction work, planned events, and unplanned traffic incidents can create massive traffic tie-ups and can have equally large economic and environmental regional impacts. With the availability of various traffic data (real-time and historically archived), how to minimize incident-induced disruption to commuting traffic and its impact to the environment presents a big challenge to global cities and communities. While unplanned incidents require careful evaluation of traffic management response plans, guidelines to develop efficient response plans for real-time operations are often lacking. Consequently, there is a real need to study unplanned traffic incidents to understand human behavior under those incidents, and learn valuable lessons to prepare public agencies to deal more effectively with large routine highway maintenance, reconstruction, big sports events, catastrophic vehicle crash and emergency situations.
Research approach
The DODE is essentially a general process of tunning parameters related to characteristics of demand and supply in real time, under general system interventions. The DODE framework integrates real-time network simulation and dynamic origin-destination estimation (DODE) algorithms, which will allow us to predict traffic impacts of futuristic pricing under real-time traffic conditions. General incidents (e.g., lane closures resulting in flow capacity drop.) will be encoded as impacts on network supply or travel behavioral change. DODE will be executed in real time, with traffic flows and speed observed in the past p minutes, and predicted by data-driven traffic prediction model in the next h minutes, to identify changes in network supply/demand caused by real-time traffic progression using a data-driven DODE framework, and to be updated every few minutes in real time.
Travel behavior in route choices under incidents will be modeled based on a disutility function for individuals and the calibrated regional network model. We will design a methodology to simulate the traffic and estimate dynamic O-D demand on the real time basis. The simulation adopts the historical traffic demand as an initial (base) demand and their pre-scribed route choices from the dynamic network model in the existing meso-scopic simulation tool (MAC-POSTS).
In the off-line manner, we first use all the data in 2018 and 2019 to learn the disutility function for individuals under incidents. Those initial disutility functions can be seen as the expected route choices of individual travelers under incidents, which can be further tuned and refined in real time provided with real-time data. Furthermore, the real-time simulation and DODE receive real-time traffic data feeds (INRIX or GPS traces) and calibrate the en-route route choices in the real time, corrects the forecast of incident-induced traffic congestion in the next hour, and computes the optimal traffic diversion ratios for pre-determined detour routes.
Schedule
This work is expected to be completed by 06/30/2022.
Timeline
Task 1. To enhance OD estimation with individual trajectories --Reduce the error between real and simulation for OD demand without incidents -- 01/01/2023 - 03/31/2023
Task.2 To develop route choice model of agents under incident with individual level. 04/01/2023 - 06/30/2023
Strategic Description / RD&T
Deployment Plan
The models and modeling framework will be integrated to a mesoscopic simulation tool that is essential to a digital twin platform for smart cities. We will work with Fujitsu North America to develop this digital twin for future deployment in U.S. cities.
Expected Outcomes/Impacts
Goodness of fit to large-scale multi-source data >= 70%
Open source all modeling implementations on Github.
Expected Outputs
TRID
Individuals Involved
| Email |
Name |
Affiliation |
Role |
Position |
| lgraff@andrew.cmu.edu |
Graff, Lindsay |
CMU |
Other |
Student - PhD |
| seanqian@cmu.edu |
Qian, Sean |
CMU |
PI |
Faculty - Tenured |
Budget
Amount of UTC Funds Awarded
$40000.00
Total Project Budget (from all funding sources)
$129677.00
Documents
Match Sources
No match sources!
Partners
| Name |
Type |
| Fujitsu North America |
Deployment Partner Deployment Partner |
