Real-Time Traffic Density Estimation System Based on Deep Learning and Big Data Technology

Traffic congestion has been increasing in large cities, affecting the environment and delaying people’s daily activities. Therefore, developing an intelligent transportation system (ITS) capable of controlling traffic flow automatically and quickly is necessary. (It is critical to develop an intelligent traffic flow that can operate automatically and quickly). With the popularity of smartphones today, GPS data on these devices is a valuable source of traffic data for traffic operations and management. Taking this as motivation, this work aims at developing a real-time data stream processing model for forecasting vehicle traffic. The research helps find the most suitable model for real traffic data through experimentation with three models: ARIMA, LSTM, Time Series Transformer. After finding a suitable model, we will train and apply the model to predict real-time traffic data. The results show that the LSTM model has the best performance with the fastest running time. Simulate a real-life system using a distributed streaming platform called Kafka to stream data every hour.