Calibration-free Traffic Signal Control Method Using Machine Learning Approaches

Liang Zhang, Wei Hua Lin

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Scopus citations

Abstract

Many existing traffic signal control strategies are operated with data from roadside surveillance systems. In recent years, vehicle-based data have become more and more accessible for various applications. In this paper, we propose a calibration-free traffic signal control scheme using vehicle-based data as input. Traffic conditions are characterized as discrete queue cycle state (DQCS) which are then used as input to the calibration-free traffic signal control scheme with the reinforcement learning approach. The k-nearest neighbor algorithm is applied in our calibration-free model. The effectiveness of the proposed model is examined with different traffic scenarios.

Original languageEnglish (US)
Title of host publicationInternational Conference on Electrical, Computer, and Energy Technologies, ICECET 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665470872
DOIs
StatePublished - 2022
Event2022 IEEE International Conference on Electrical, Computer, and Energy Technologies, ICECET 2022 - Prague, Czech Republic
Duration: Jul 20 2022Jul 22 2022

Publication series

NameInternational Conference on Electrical, Computer, and Energy Technologies, ICECET 2022

Conference

Conference2022 IEEE International Conference on Electrical, Computer, and Energy Technologies, ICECET 2022
Country/TerritoryCzech Republic
CityPrague
Period7/20/227/22/22

Keywords

  • calibration-free model
  • machine learning
  • queue length
  • traffic signal control

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Computer Vision and Pattern Recognition
  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering
  • Media Technology
  • Instrumentation

Fingerprint

Dive into the research topics of 'Calibration-free Traffic Signal Control Method Using Machine Learning Approaches'. Together they form a unique fingerprint.

Cite this