SDN Control-Enabled and Time-Quantum-Driven Max-Pressure Approach for Intersection Management in Smart City

Abstract

Congestion optimization, at an intersection, needs real-time traffic state information from the neighboring intersections. Existing actuated traffic light controllers (TLCs) act on local traffic data. On the same line, max-pressure-based TLCs have been proved effective to act in real-time. However, such TLCs act only on queue length or travel time, which are unable to optimize the congestion as they act only on local traffic data. Furthermore, such TLCs suffer from the starvation problem; if a road lane, at an intersection, continuously receives enormous vehicles during peak hours, then it gets continuous green phase duration, and other lanes continue with red phase duration for a longer time. To overcome this, we propose an efficient quantized max-pressure-based TLC utilizing software-defined networking (SDN) to mitigate the abovementioned issues and maintain the traffic flow in the city. The proposed SDN-enabled quantized max-pressure controller (SDN-QMPC) takes real-time vehicular dynamics as inputs and generates the traffic light phase duration in the form of dynamic <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">time quantum</i> by accounting for heterogeneous traffic states. Roadside units/cameras are considered vehicle sensing entities for the real-life implementations to account vehicular dynamics, such as vehicle speed, vehicle position, etc. These parameters become inputs for SDN-QMPC to generate SDN control-enabled TLC signals. An extensive simulation study has been carried out on an Indian city map using an open-source simulator, i.e., simulation of urban mobility. The comparative results against the state-of-the-models prove the effectiveness of SDN-QMPC.