Extending Intelligent Traffic Control with Deep Q-Learning — Work in Progress
A short update on the control-side continuation of my M.S. thesis, extending the Deep Q-Learning traffic-signal-control research before submission for peer review.
My 2021 M.S. thesis at Bu-Ali Sina University focused on adaptive traffic-signal control with Reinforcement Learning and Deep Q-Learning. The work modeled a four-way intersection in SUMO, represented traffic state for sequential decision-making, and used a neural-network-based Q-learning agent to select signal phases according to traffic conditions.
Current research continuation
The control side of that research is now being developed into a new manuscript. The new work remains part of the Deep Q-Learning traffic-control line, with an updated formulation and broader evaluation being prepared for peer review.
Because the manuscript is not yet published, I am intentionally keeping unpublished methodological details and full result tables off the public website.
Research lineage
This work is part of one continuous research path:
M.S. Deep Q-Learning traffic-control thesis → vehicle perception → public vehicle datasets → published object-detection papers → new traffic-control manuscript
After submission and peer review, this page can be updated with the final bibliographic record and the technical details that are appropriate to publish.