Deploying Traffic Smoothing Cruise Controllers Learned from Trajectory Data

Lichtle, Nathan; Vinitsky, Eugene; Nice, Matthew; Seibold, Benjamin; Work, Daniel B.; Bayen, Alexandre M.

doi:10.1109/icra46639.2022.9811912

Cited by 16 publications

(1 citation statement)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, it involves planning a fleet of AVs to accomplish a given objective at the system level, which involves both AVs and human drivers. These objectives include mitigating traffic congestion [1], curbing emissions [2], and promoting smoother traffic flows [3].…”

Section: Introductionmentioning

confidence: 99%

Learning Eco-Driving Strategies at Signalized Intersections

Jayawardana¹,

Wu²

2022

2022 European Control Conference (ECC)

View full text Add to dashboard Cite

Conventional control, such as model-based control, is commonly utilized in autonomous driving due to its efficiency and reliability. However, real-world autonomous driving contends with a multitude of diverse traffic scenarios that are challenging for these planning algorithms. Modelfree Deep Reinforcement Learning (DRL) presents a promising avenue in this direction, but learning DRL control policies that generalize to multiple traffic scenarios is still a challenge. To address this, we introduce Multi-residual Task Learning (MRTL), a generic learning framework based on multi-task learning that, for a set of task scenarios, decomposes the control into nominal components that are effectively solved by conventional control methods and residual terms which are solved using learning. We employ MRTL for fleet-level emission reduction in mixed traffic using autonomous vehicles as a means of system control. By analyzing the performance of MRTL across nearly 600 signalized intersections and 1200 traffic scenarios, we demonstrate that it emerges as a promising approach to synergize the strengths of DRL and conventional methods in generalizable control.

show abstract

Section: Introductionmentioning

confidence: 99%