With the rapid development of China’s economy, traffic congestion has become a serious problem affecting the efficiency and safety of the traffic system, especially in urban regions with high traffic density. Due to the lack of effective forecasting methods, traffic congestion events seriously affect normal operation of the intracity traffic network. In order to achieve better prediction results, a type of gated recurrent neural network—long short-term memory neural networks—were used to build the model. The prediction accuracies for different tasks all approach 85%. Then, several different factors which may influence the congestion prediction were analyzed to find why LSTM could not fit the congestion change better. In order to have a comprehensive understanding of the model based on the LSTMs, several algorithms were studied by building models. As the result, the prediction accuracies of these new models are noticeably lower than those of the LSTM models.
Prediction of Distribution of Traffic Congestion on High Traffic Density Region Based on Deep Learning
19th COTA International Conference of Transportation Professionals ; 2019 ; Nanjing, China
CICTP 2019 ; 2211-2223
2019-07-02
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
TRAFFIC CONGESTION PREDICTION DEVICE AND TRAFFIC CONGESTION PREDICTION METHOD
Europäisches Patentamt | 2017
|Traffic congestion prediction method and traffic congestion prediction device
Europäisches Patentamt | 2016
|TRAFFIC CONGESTION PREDICTION APPARATUS AND TRAFFIC CONGESTION PREDICTION PROGRAM
Europäisches Patentamt | 2015
|TRAFFIC-CONGESTION PREDICTION SYSTEM, TRAFFIC-CONGESTION PREDICTION METHOD, AND RECORDING MEDIUM
Europäisches Patentamt | 2017
|Europäisches Patentamt | 2019
|