An Adaptive and Coordinated Traffic Signal Scheme for Greener Transport 4.0

Abstract

AbstractDue to increased population and urbanization, traffic congestion has become a major concern. In this research, we introduce an adaptive traffic signaling scheme based on road traffic density to facilitate optimal traffic signal control as well as effective traffic management. The proposed scheme uses live video as an input provided to a deep Q network to provide adaptive phase timings as the output. Compared to the existing works, we introduced per car unit (PCU) as a novel input to represent the effect of each vehicle type on traffic conditions. Extensive tests on real-time data amply prove that the proposed scheme enhances the average speed of traffic up to 5.597 km/h. The proposed scheme shows an average increment of 175.71% in average mean speed compared to the existing static schemes. Except for the high traffic scenario, for both mid-traffic and low-traffic scenarios, the proposed scheme shows a considerable improvement in both average densities and maximum densities. In the mid-traffic scenario, the average speed shows an improvement of 3.85 km/h, while in the low-traffic scenario, the average mean speed shows an improvement of 7.96 km/h. A reduction in fuel consumption and average delay were also observed. A reduction in fuel consumption and average delay also observed, which will lead to a greener Transport 4.0.KeywordsTraffic controlVideo processingQ-learningAdaptiveCoordinated traffic signalingPer Car Unit (PCU)