100th TRB Annual Meeting

Online
January 21-29

100th TRB Annual Meeting

Once again, Aimsun is the proud sponsor of the app for TRB AM – a sponsorship that is doubly important for this virtual celebration of the TRB centenary annual meeting. The meeting program will cover all transportation modes, addressing topics of interest to policy makers, administrators, practitioners, researchers, and representatives of government, industry, and academic institutions. A number of sessions and workshops will focus on the spotlight theme for the 2021 meeting: Launching a New Century of Mobility and Quality of Life. This year’s meeting also will feature dozens of sessions on how COVID-19 has impacted transportation.

Look out for:

Continual Learning of Microscopic Traffic Models using Neural Networks

Poster Session

1061 – Traffic Flow Theory and Characteristics, Part 1: Emerging Trends in Traffic Flow Theory: Connected and Automated Vehicles and Data-driven Methods

Date and time

Monday, January 25 10:00 AM- 11:30 AM ET

Presentation number

TRBAM-21-03666

Authors

Yashar Farid, Toyota InfoTech Labs
Abdul Rahman Kreidieh, University of California, Berkeley
Farnoush Khalighi, Aimsun, Inc
Hans Lobel, Pontificia Universidad Catolica de Chile
Alexandre Bayen, University of California, Berkeley

Abstract

In a mixed-autonomy traffic scenario, where human drivers and autonomous vehicles share the streets, self-driving cars need to be able to predict in a robust manner the behavior of human-driven vehicles, in order to guarantee a safe and smooth driving experience. Although traffic theory provides several models of human drivers, these models are often parameterized by few parameters which can limit their performance in modeling complex behaviors. The lack of sufficient model capacity and the behavioral shifts in human driving reduces the usefulness of these methods in real-life situations. Based on recent advances in trust region optimization, we present a new method for data-efficient continual learning, that allows to incrementally train a high-performance driver model, while avoiding the effects of catastrophic forgetting. The proposed approach focuses on keeping output distributions of previous tasks stable during training in new scenarios, by using explicit constraints in the optimization problem based on Kullback–Leibler divergence. As a result, we observe minimal loss of performance in previous tasks, while increasing the generalization capabilities of the learned representations. We evaluate the performance of the proposed method in traffic modeling tasks, including mandatory lane change and acceleration tasks. Vehicle trajectory data from Next Generation Simulation (NGSIM) is used for training and validation of the models. Results show state-of-the-art performance in the presence of scenarios with small amounts of data.

Assessment of the road network resiliency in the presence of road maintenance sites: an environmental perspective

Poster session

1225 – Innovations in Roadside Maintenance

Date and time

Tuesday, January 26 2:30 PM- 4:00 PM ET

Authors

Behzad Bamdad Mehrabani, Université Catholique de Louvain
Kaitlin Stack Whitney, Rochester Institute of Technology (RIT)

Presentation number

P21-20464

Abstract

This study, the resilience of the road network (from an environmental perspective), in the presence of road maintenance sites, investigated the environmental degradation caused by additional delay in road maintenance sites. Traffic simulation in Aimsun Next software was implemented, and the Sioux Falls road network selected as the case study. The closure of one lane in each section represent the presence of road maintenance site. The results of this study show that the existence of road maintenance sites has a significant impact on the production of pollutants in the network. The higher the percentage of sections in which a lane is closed, the lower the resilience of the network.

TRB AM webpage Online programme


Related topics

Events Articles Latest