Smart Traffic Management for Urban Areas: An IoT-Driven GS-SVM Approach for Real-Time Congestion and Safety Control

In many countries, especially India, the sum of automobiles in urban areas has augmented dramatically in recent years. Due to the current infrastructure and the rapidly growing number of vehicles, traffic systems are at a standstill during peak hours. The movement of overloaded vehicles outside of their designated zone and time, careless driving, and disregard for traffic safety regulations are some of the major issues facing traffic management. This paper suggests a real-time Intelligent Traffic Signal System (ITSS) based on the Internet of Things (IoT). It uses a programmable micro-controller and inductive loops to measure traffic density. For seamless vehicle mobility with minimal delay, the centralised control unit's inter-communication sets the traffic light's timer and synchronises with the traffic density in real-time. As part of the development of smart cities, this study focusses on creating an appropriate strategy to lower traffic accidents and enhance individual traffic management. To gather information on traffic patterns besides direct cars along the most efficient routes, a modified machine learning-based model that makes use of a Support Vector Machine (SVM) is suggested. Approximately 8,000 roadside 12-hour manual counts were gathered from Kaggle. The Highway Traffic Crash Dataset is the source of the information. Information is extracted from the gathered data using statistical features, correlation-based features, higher-order statistical features, and database features. The Binary Grey Wolf Optimiser (BGWO) is used to identify the best features for classification from the extracted data. Next, SVM is used to forecast traffic congestion data. Grid Search is used to fine-tune the model, and the suggested model is called GS-SVM. Associated to other smart city traffic management mechanisms, the experimental authentication demonstrates better traffic control in smart cities.