Intelligent traffic management system based on historical data analysis

Abstract

The problem of vehicular traffic congestion is ubiquitous yet non-trivial. It is increasingly worsening by the day all around the world with severe vehicular traffic taking its toll on all road users. With the upsurge in urban traffic jams, innovative control strategies are therefore essential to allow efficient flow of vehicular movement. It is thus not surprising that a myriad of novel control strategies has been developed over the past years to manage the ever-growing urban gridlock. Many of the currently used traffic control strategies are based on the relatively inefficient fixed-time traffic systems, like in the case of Ghana, or on a central traffic-responsive control system, which is challenging to implement and even much more difficult to maintain. As a consequence of inefficiencies in traffic control, road users are saddled with inconveniently longer waiting times in queues. To mitigate this problem, we proposed a distributed artificial intelligence and multi-agent system as a viable approach to manage the traffic menace. The proposed system uses historical data for traffic management and was designed and implemented using Simulation of Urban Mobility (SUMO) software. The result obtained in the comparison of the current fixed time-controlled system and designed system clearly indicated that the proposed system outperformed the fixed-time cycle controllers in every key performance index selected for evaluation.