Enhancing Urban Traffic Management through Predictive Modelling and Drone-Captured Image Analysis for Smart Traffic Lights

This research paper explores the utilization of predictive modelling and drone-captured image analysis to enhance urban traffic management in the context of smart traffic lights. The study focuses on employing advanced machine learning techniques, including LSTM and GRU architectures, to predict traffic flow patterns. Comparative analysis is conducted by evaluating the performance of these deep learning models against traditional algorithms such as Linear Regression, Gradient Boosting Regressor, and Random Forest Regressor. Metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R-squared values are utilized to quantify the predictive accuracy of these models. Experimental results reveal that the LSTM model achieves an MAE of 6.32 and an RMSE of 12.76, while the GRU model yields an MAE of 6.50 and an RMSE of 13.12. These values outperform traditional algorithms, emphasizing the effectiveness of the proposed models in improving traffic flow predictions. The dataset comprises drone-captured images of urban traffic scenes, enabling the extraction of relevant features for accurate predictions. Findings underscore the potential of the proposed models in advancing intelligent traffic management systems.