Continuous expansion of cities has resulted in the growing traffic network to cater the demand of increasing number of vehicles. The rapid growth in population, unplanned development, and inadequate infrastructure have led to several problems like pollution and traffic congestion. Traffic congestion is one such issue with which not only metropolitan cities but also medium and small cities are dealing on day-to-day basis. Intelligent transportation system (ITS) helps in providing relief to the traffic congestion-related problems. Accurate prediction of short-term traffic flow is an important pillar of ITS. Several researchers have tried to model and predict traffic flow using different methods/techniques. Heterogeneous traffic flow makes traffic studies often more critical and challenging. This study aims to find the optimal deep bidirectional long short-term memory (LSTM) neural network to predict the short-term traffic flow under heterogeneous traffic conditions. Real data from an urban location in Delhi was collected using video cameras. Deep learning techniques like LSTM neural network and bidirectional long short-term memory (Bi-LSTM) neural network (NN) were used in this study to predict the traffic flow for the next 5 min using collected data. It was found that Bi-LSTM neural network with 2 layers gives better results as compared to LSTM NN, single–layer Bi-LSTM NN, and three-layered Bi-LSTM NN.
Deep Bi-LSTM Neural Network for Short-Term Traffic Flow Prediction Under Heterogeneous Traffic Conditions
Lecture Notes in Civil Engineering
Recent Advances in Transportation Systems Engineering and Management ; Chapter : 39 ; 597-611
2022-11-11
15 pages
Article/Chapter (Book)
Electronic Resource
English
Short-term traffic flow prediction method based on improved LSTM
European Patent Office | 2021
|