Real-time traffic signal optimization for urban mobility: a reinforcement learning-enhanced framework with...

Area of Science:

• Intelligent Transportation Systems
• Computer Engineering
• Environmental Science

Background:

• Urban mobility faces challenges from increasing traffic congestion and associated environmental impacts.
• Smart city initiatives require efficient and adaptable traffic management solutions.
• Existing traffic control systems often struggle with dynamic conditions and energy constraints.

Purpose of the Study:

• To develop an intelligent and adaptable traffic control strategy for enhancing urban mobility in smart cities.
• To minimize traffic wait times, reduce congestion, and improve environmental health.
• To evaluate the effectiveness of various traffic management algorithms and hardware implementations.

Main Methods:

• Investigated and evaluated rule-based, optimization-based, and machine learning (Reinforcement Learning) algorithms.
• Employed microscopic traffic simulations and statistical analyses for performance evaluation.
• Implemented algorithms on Field Programmable Gate Array (FPGA) platforms for real-time processing.

Main Results:

• The proposed system demonstrated significant improvements in traffic flow and congestion reduction.
• FPGA implementation achieved over 7x speedup compared to general-purpose processing units (GPPUs).
• The system effectively reduces fuel consumption and carbon dioxide emissions.

Conclusions:

• The intelligent, adaptable traffic control strategy enhances urban mobility and environmental quality.
• FPGA-based implementation offers efficient, low-latency processing for smart city traffic management.
• The solution provides a viable approach to address traffic challenges and improve air quality, particularly in regions like Kuwait.