International Journal of Advanced Research in Computer and Communication Engineering

Abstract: Urban traffic congestion severely affects emergency vehicle response times, leading to preventable loss of life and reduced efficiency of public safety systems. Traditional traffic management relies on static signal plans and fixed routing, which cannot adapt to sudden congestion, roadblocks, or peak-hour variations. Recent research on Multi-Agent Systems (MAS), Reinforcement Learning (RL), and vehicle-to-infrastructure (V2I) coordination has shown promising results for dynamic and intelligent emergency mobility. This review analyzes ten influential studies across three domains: emergency vehicle routing algorithms, learning-based traffic signal control, and cooperative multi-agent negotiation frameworks. Prior work such as EMVLight and MARL-based traffic control demonstrates that decentralized agents can learn to reduce delays, but most systems are limited to small grids, isolated intersections, or single-agent routing. Few provide city-scale simulations integrating ambulances, fire trucks, and traffic lights within a shared communication environment. To address these gaps, this paper highlights the need for a unified, scalable simulation combining SUMO traffic modeling with SPADE-based agent communication, enabling adaptive routing, green-wave negotiation, and real-time scenario testing. The review concludes that multi-agent simulation is a practical and scalable approach for optimizing emergency response in future smart cities.

Keywords: Multi-Agent Systems, SUMO, SPADE, Reinforcement Learning, Emergency Vehicle Routing, Traffic Signal Control, Vehicle-to-Infrastructure Communication, Intelligent Transportation Systems, Q-Learning, Smart Cities

Downloads: | DOI: 10.17148/IJARCCE.2025.141102