Addressing the optimization of coordinated control of traffic signals at intersections within road traffic networks, this paper proposes a deep reinforcement learning algorithm based on D3QN to enhance the efficiency of intersection passage and alleviate traffic congestion. By delving into the underlying implicit features contained within traffic state information [1], the algorithm empowers intelligent intersection signal control units with the capability to autonomously learn based on traffic conditions. It utilizes a deep neural network to learn traffic state information and predict the impact of different control strategies on future traffic flows, thereby selecting the optimal signal control strategy. The D3QN algorithm estimates the value function, taking the current state as input and providing expected returns for each action. It selects actions to maximize long-term rewards, enhancing intersection signal efficiency and network optimization. Experiments show it can quickly stabilize and optimize intersection signal control, with good results in traffic networks.
Traffic Signal Timing Optimization Based on D3QLight
Lect. Notes Electrical Eng.
International Conference on Artificial Intelligence and Autonomous Transportation ; 2024 ; Beijing, China December 06, 2024 - December 08, 2024
The Proceedings of 2024 International Conference on Artificial Intelligence and Autonomous Transportation ; Chapter : 28 ; 259-267
2025-03-16
9 pages
Article/Chapter (Book)
Electronic Resource
English
Traffic Signal Timing Optimization
Transportation Research Record | 2013
|Traffic Signal Timing Optimization Based on Genetic Algorithm
Springer Verlag | 2022
|