Traffic Flow Prediction for Reducing Urban Congestion using Improved Bidirectional Gated Recurrent Unit

Urbanization is a global trend, with more people than ever residing in big cities. This surge in urban populations has created numerous challenges, one of the most pressing being traffic urban congestion. However, lack of transparent explanations for traffic flow in urban congestion faces challenges in detecting traffic flow patterns and attacks caused insufficient traffic management. This paper, proposed Improved Bidirectional Gated Recurrent Units (Improved Bi-GRU) is efficiently selecting the essential feature from high-dimensional Internet of Things network traffic data. The Bi-GRU ensure that the most impactful features are used for temporal and contextual analysis which improve classification accuracy. Initially data obtained from the PEMS04 dataset and pre-processing using Wavelet-based multiscale denoising technique effectively captures both short-term fluctuations and long-term trends in traffic flow data by decomposing it into multiple time-frequency components. Detection using improved Bi-GRU ability to understand temporal and contextual relationship helps to detecting subtle deviations in traffic patterns indicative of such attacks. The proposed method achieves better Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) compared with existing method such as Convolutional Neural Network (CNN) – Recurrent Neural Network (RNN) network.