S-Edge: Smart Edge Computing Framework for Real-time Heterogeneous Vehicular Network

Abstract

With the rapid growth of transportation vehicles, urban centers are becoming overcrowded due to limited road infrastructure. Several queue length-based traffic light controllers have been developed to address this problem. Due to excessive congestion on the road during peak hours, the existing system suffers from the starvation problem at any intersection. This results in numerous instances where longer green phase duration is assigned to the same lane, increasing vehicle waiting time in other lanes. This issue is addressed by an efficient Smart Edge (S-Edge) lane pressure-based traffic light controller framework that accounts for the real-time heterogeneous vehicular dynamics. Additionally, this work proposes a method that uses average queue length and waiting time to estimate lane pressure for the Edge-controller that allocates phase duration effectively. This light-weighted actuated traffic light controller determines the cycle and phase (R/Y/G) durations of traffic lights. To validate the effectiveness of the proposed S-Edge controller, a detailed analysis has been carried out against the same line of state-of-the-art models that are based on a well-known open-source simulator called Simulation of Urban MObility (SUMO).