• Classified by Topic • Classified by Publication Type • Sorted by Date • Sorted by First Author Last Name • Classified by Funding Source •
Learning a robust multiagent driving policy for traffic congestion reduction.
Yulin
Zhang, William Macke, Jiaxun Cui,
Sharon Hornstein, Daniel Urieli, and Peter
Stone.
Neural Computing and Applications, 2023.
In most modern cities, traffic congestion is one of the most salient societalchallenges. Past research has shown that inserting a limited number ofautonomous vehicles (AVs) within the traffic flow, with driving policies learnedspecifically for the purpose of reducing congestion, can significantly improvetraffic conditions. However, to date, these AV policies have generally beenevaluated under the same limited conditions under which they were trained. Onthe other hand, to be considered for practical deployment, they must be robustto a wide variety of traffic conditions. This article establishes for the firsttime that a multiagent driving policy can be trained in such a way that itgeneralizes to different traffic flows, AV penetration, and road geometries,including on multilane roads. Inspired by our successful results in ahigh-fidelity microsimulation, this article further contributes a novelextension of the well-known cell transmission model (CTM) that, unlike the pastCTMs, is suitable for modeling congestion in traffic networks, and is thussuitable for studying congestion reduction policies such as those considered in
@Article{yulin_zhang_NCAA2023,
author = {Yulin Zhang and William Macke and Jiaxun Cui and Sharon Hornstein and Daniel Urieli and Peter Stone},
title = {Learning a robust multiagent driving policy for traffic congestion reduction},
journal = {Neural Computing and Applications},
year = {2023},
abstract = {
In most modern cities, traffic congestion is one of the most salient societal
challenges. Past research has shown that inserting a limited number of
autonomous vehicles (AVs) within the traffic flow, with driving policies learned
specifically for the purpose of reducing congestion, can significantly improve
traffic conditions. However, to date, these AV policies have generally been
evaluated under the same limited conditions under which they were trained. On
the other hand, to be considered for practical deployment, they must be robust
to a wide variety of traffic conditions. This article establishes for the first
time that a multiagent driving policy can be trained in such a way that it
generalizes to different traffic flows, AV penetration, and road geometries,
including on multilane roads. Inspired by our successful results in a
high-fidelity microsimulation, this article further contributes a novel
extension of the well-known cell transmission model (CTM) that, unlike the past
CTMs, is suitable for modeling congestion in traffic networks, and is thus
suitable for studying congestion reduction policies such as those considered in
},
}
Generated by bib2html.pl (written by Patrick Riley ) on Wed Aug 14, 2024 12:58:46