Machine Learning Algorithm Comparison for Traffic Signal: A Design Approach

Abstract

The number of vehicles is increasing significantly every day, especially in major cities. To control traffic flow on extensive road networks and facilitate crossing traffic flow on extensive road networks and facilitate crossing crossings, intersection n traffic signal coordination is required. It is therefore a new technology in the present world for intelligent traffic control and administration systems due to the number of vehicles with so many communication restrictions in arrays. The traffic signal strives to boost the effectiveness of transportation networks, increase safety, and reduce congestion. Technological advances enable better monitoring and management of traffic flow, and traffic signal design and operation are continually developing. The employment of some algorithms for traffic light control has improved the intelligence of traffic light controllers. This research study has developed data instruction strategies using data mining and image processing techniques. By comparing the analysis of several machine learning algorithms for predicting traffic flow and patterns, such as KNN, SVM, DNN, and Random Forest, the accuracy can be increased by lowering the overfitting.