Deep Reinforcement Learning–Based Adaptive Traffic Signal Optimization System For Urban Congestion
P. Neela Sundari,
Yenuganti Maheshwari,
Yemineni Shiva Pujitha,
Vasantha Manisha,
Yarrakula Kavya Sri,
Vutti Varshitha,
Nakka Vijaya Lakshmi
The proposed system, an Adaptive Traffic Signal Control System, utilizes Deep Reinforcement Learning
to dynamically adjust traffic signal timings, minimizing vehicle waiting time and congestion, and
optimizing traffic flow in real-time. This approach addresses the limitations of traditional fixed-timing
systems, which are inefficient under dynamic traffic conditions. By learning optimal signal policies
through interaction with a simulated traffic environment, the system aims to reduce traffic congestion
and waiting time, improving overall traffic efficiency. The significance of this system lies in its ability to
enable intelligent decision-making in smart cities, providing a scalable and real-time solution for urban
congestion control. The system’s effectiveness is crucial in reducing travel time, fuel consumption, and
pollution, ultimately enhancing the quality of life for urban residents.
