Jordan Journal of Civil Engineering

Reinforcement Learning in Urban Network Traffic Signal Control


Eslam AL-kharabsheh;


Traffic signal recognition and anticipation are essential for both advanced driver assistance systems; due to their superior performance in data categorization, deep learning has gained significance in vision-based object identification in recent years. The color space of the input image and the deep learning network model are reviewed as two of the system's main components when examining the application of deep learning to the creation of a high-performance traffic signal detection system in Urban. Using distinct network models based on the Faster R-CNN algorithm and color spaces, the RGB color space and Faster R-CNN model perform exceptionally well in simulations; these data may be used to develop a system for traffic signal detection; the majority of engineers are developing a new traffic signal that requires image recognition. The result is taught to a connected softmax and linear regression layer, which classifies and outputs the bounding boxes of objects.


Bounding Boxes, Faster R-CNN, Modelled Environments, Simulation, Traffic Signal Detecting System