RPG Seminar – Group and Lead: Intelligent Vehicle Control for Traffic Congestion Alleviation at Intersections
Date:
18 May, 2021 (Tue)
Time:
4:00 pm
Webinar Link:
https://hku.zoom.us/j/92307999396?pwd=YWFabmNGbkVNSWhvMWFkVWQvdi81dz09

Add to your calendar: iCal, Google Calendar

HKU Campus Map


Speaker

Ms. Miaomiao CAO,
Department of Electrical and Electronic Engineering
The University of Hong Kong

Abstract

Intersections are prone to congestion in urban cities and various methods of traffic signal control have been designed to reduce congestion. At the same time, making competent speed plans for vehicles around the intersection to efficiently utilize green phase resources is also significant for congestion alleviation. The key is to improve the opportunities of vehicles to pass the intersection within the green time, and meanwhile reduce their waiting time and queue length at red lights. In this work, we propose a novel DeepGAL model for vehicle control, dividing vehicles into different groups, assigning the leaders in each group and delivering intelligent control on leaders with deep reinforcement learning. Extensive experiments are conducted based on two real-world datasets with distinct traffic flow rates, and the results demonstrate that our DeepGAL model achieves outstanding improvement over various performance metrics compared with five classic car-following models. We also test our DeepGAL model under indeterministic traffic signal phase and timing (SPAT) information, which indicates the effectiveness of our method even with partial SPAT information. In addition, considering the practical issue that only part of the vehicles could be controlled by our scheme, we conduct simulations on different penetration levels, which demonstrates that even with only 10% penetration of leader candidates, our method can greatly alleviate traffic congestion at intersections, which validates the feasibility of our DeepGAL model in practice.

 

Zoom Link (during COVID-19 special period):
https://hku.zoom.us/j/92307999396?pwd=YWFabmNGbkVNSWhvMWFkVWQvdi81dz09

Meeting ID: 923 0799 9396
Password: 758624

 

Biography of the speaker:
CAO Miaomiao received the B.Eng. degree in Electric Power Engineering and Automation from Shanghai Jiao Tong University, China, in 2012 and the MSc degree in Electrical and Electronic Engineering from University of Hong Kong, Hong Kong, China, in 2016. She is currently a PhD student at the Department of Electrical and Electronic Engineering of University of Hong Kong. Her research interests include traffic signal control, deep reinforcement learning and intelligent transportation systems.

 

All are welcome!


Organizer

Prof. V.O.K. Li

Most seminars are open to the general public, free of charge, unless otherwise stated. Registration is not required. Arrangement for car parking facilities on campus please contact us for details.

For enquiries, please contact:
Department of Electrical and Electronic Engineering,
Room 601, Chow Yei Ching Building,
Pokfulam Road, Hong Kong
Tel: (852) 3917 7093
Email: seminar@eee.hku.hk