Emergency vehicle signal priority control method for arterial intersections in an intelligent connected environment

Area of Science:

• Intelligent Transportation Systems
• Traffic Signal Control
• Reinforcement Learning

Background:

• Emergency vehicle (EV) progression is hindered by traffic congestion and signal coordination issues in connected environments.
• Existing methods struggle with upstream queues, downstream spillback, and dynamic traffic conditions.
• Need for advanced control strategies to optimize EV passage through multiple intersections.

Purpose of the Study:

• To propose a novel Spatiotemporal Graph Multi-Agent Proximal Policy Optimization (STG-MAPPO) method for enhanced emergency vehicle signal priority control.
• To improve emergency vehicle progression efficiency and reduce overall traffic delays in intelligent connected environments.
• To explicitly model heterogeneous yielding behaviors of connected and automated vehicles (CAVs) and human-driven vehicles (HDVs).

Main Methods:

• Developed STG-MAPPO by modeling intersections as cooperative agents and incorporating EV arrival prediction, yielding capacity, and blockage risk.
• Utilized a spatiotemporal graph encoder to capture topological coupling and traffic-state propagation among intersections.
• Integrated heterogeneous yielding responses of CAVs and HDVs into the target-lane yielding model.

Main Results:

• STG-MAPPO demonstrated superior performance in EV progression efficiency, reduced general traffic delay, and improved safety indicators compared to baseline methods.
• Simulations showed significant reductions in EV travel time (up to 22.28%), EV delay (up to 51.02%), and EV stop frequency (up to 60.00%) compared to DDQN.
• Compared to MAPPO, STG-MAPPO achieved reductions of 8.54% in EV travel time, 22.58% in EV delay, and 33.33% in EV stop frequency.

Conclusions:

• The proposed STG-MAPPO method effectively addresses challenges in emergency vehicle signal priority control within intelligent connected environments.
• This approach offers substantial improvements in traffic flow efficiency, delay reduction, and safety for emergency vehicles.
• The method's ability to handle complex traffic dynamics and heterogeneous vehicle behaviors makes it a promising solution for future intelligent transportation systems.