Efficient coordination in agent based traffic control and management

Abstract

This paper presents a new multi agent based approach applying computational intelligence concepts to implement a cooperative and well coordinated, Multiagent system for real-time traffic control and management of a complex traffic network. The large-scale traffic network is divided into various sub problems, and each sub problem is handled by an intelligent agent with fuzzy neural and Case based reasoning based decision-making module.Proposed solution help in quick incident diagnosis, minimizing response delays by managing them and speed up the traffic incident recovery to avoid the congestion. Key to effective response to any incident requires extensive communication and institutional coordination .Ontology is used as the base component for coordination and collaboration to address the research issues of efficient coordination and consensus development among multiple autonomous agents acting as robots or virtual cops at various intersections for decision making. The architecture counts on several AI techniques germane to efficient decision making i.e. CBR & FNN. The proposed architecture shows a high degree of adaptability leading to the least need for human intervention. The proposed solution for traffic management is followed by a case study that reveals the performance and effectiveness of the proposed architecture over previous architectures.