Managing traffic on highways is a cumbersome task, which requires accurate vehicle type analysis, traffic jam prediction, speed analysis, direction prediction, etc. In order to perform all these tasks efficiently sensing devices are connected at highways. The connected devices process data locally and take actions based on this data processing. These actions include modifying speed zones, deploying traffic authorities, and communicating with other vehicles. In order to do these tasks efficiently, it is recommended that these systems should be able to predict and broadcast this information as quickly as possible. In order to resolve this issue, the underlying text proposes a Q‐learning inspired highway traffic management system. This system incorporates different kinds of Internet of things (IoT)‐enabled sensors for speed monitoring, vehicle type detection, collision prediction and congestion prediction for evaluating different highway parameters. These parameters are then given to a machine learning model for improving highway traffic flow via high speed and high accuracy decision making and information dispersal. It is observed that deploying the proposed system is able to improve highway traffic flow by 14%, reduce accidents by 20% and reduce road congestion by 9% when compared with standard non‐machine learning deployed systems. Improving this flow quality also results in better road quality maintenance.
Machine Learning‐Based Traffic Operation System
Urban Transportation Systems ; 245-262
2024-12-29
18 pages
Article/Chapter (Book)
Electronic Resource
English