SMART TRAFFIC CONTROL MODEL USING MACHINE LEARNING APPROACH

ABSTRACT

The fast-growing increase of vehicles in developed, developing and under-developed countries are alarming and this calls for the need to provide a smart system that will be able to control the traffic so as to curb or limit the rate of occurrence of an accident in our major roads and to pave ways for free flow of vehicles. The traditional and existing traffic congestion control system is based on the time-based scheduling approach where there is a change in colours from red to green at a particular time interval to control the vehicles. This approach or technique is not sufficient to meet the demands of the increasing number of vehicles that ply roads on daily basis. The aim of this project is to develop a machine learning model to recognize both traffic and non – traffic congestions for traffic control purposes. Two hundred Imaging Dataset of traffic jam and free roads scenarios were collected and Grey Level Co-occurrence matrix features were extracted. Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Random Forest (RF) machine learning algorithms were used for the classification of our dataset. Support Vector Machine (SVM) gave an accuracy of 80%, K-Nearest Neighbors (KNN) gave an accuracy of 90%, and Random Forest (RF) also gave an accuracy of 90%. Implementation may be carried out and also an accuracy of 90% from both Random Forest and K-Nearest Neighbors implies that more machine Learning Algorithms need to be tested.

Keywords: Intelligent traffic control, Machine learning Algorithms, Smart Traffic Light, Traffic Flow Prediction.