An Intelligent Sequential Approach for Traffic Single Agent Lights Signals Control

Traffic congestion in cities poses challenges, waiting time and frustrating commuters. Deep Reinforcement Learning (DRL) offers solutions by enabling agents to learn optimal strategies. This paper explores DRL's application in optimizing urban traffic light systems. Recent advancements in reinforcement learning, including deep neural networks and real-time data integration, promise intelligent adaptive traffic lights. This paper introduces a new intelligent sequential approach named SSE_DQN (Self Sequential Estimator Deep Q Network), this approach enhances learning performance by incorporating insights like optimal lane-changing and dynamic condition responses. SSE_DQN represents a significant advancement, potentially alleviating congestion and improving urban transportation efficiency. Through simulations and real-world testing, its effectiveness in adapting to traffic patterns is demonstrated. This research contributes to intelligent transportation systems, showcasing DRL's potential in addressing urban challenges.