Search Results

With the development of vehicle-road coordination technology, driving modes of the vehicle are in the process of development from manual driving, assisted driving, autonomous driving, mixed driving between people and vehicles to advanced unmanned driving. Heterogeneous traffic flows are essential for the development of vehicle-road coordination systems. However, in real life, it is necessary to intelligently monitor heterogeneous traffic flow because they involve many types of vehicles, complex scenarios, complex hidden factors in traffic conditions, and different operating characteristics of vehicles at different times and places.

Robust and accurate short-term traffic volume prediction methods are indispensable in driving assistance and active traffic control and management. With the popularity of deep learning, the hybrid methods play an important role in improving the prediction accuracy. To fully cover the spatial-temporal characteristics of traffic flow, this paper proposes an attention-based spatial and temporal model (AST) through combining convolutional neural network (CNN), gated recurrent unit (GRU). Besides, the attention mechanism is also introduced after GRU to further improve the prediction accuracy. The experiments which are carried on the Beijing expressway traffic volume data indicate that the AST model has better performance than the baseline models in terms of prediction accuracy. Compared with ARIMA, SVR, CNN, and GRU, the MAE of AST is reduced by 21.6%, 20.9%, 11.0%, and 9.9%; the MAPE is reduced by 14.4%, 15.1%, 10.7%, and 10.3%; the RMSE is reduced by 22.6%, 20.3%, 11.0%, and 10.8%.

At present, prediction values of accurate traffic data by highway traffic control departments are not accurate enough. To provide better traffic guidance for pedestrians, new methods must be used to estimate traffic speed data with less error. This paper proposes an attention-convolution-bidirectional long short-term memory model that considers both temporal and spatial factors, combining a convolutional neural network with spatial local feature extraction capabilities and a bidirectional long short-term memory that can simultaneously consider long-term information in the forward and backward directions. Then add a layer of attention mechanism at the top to make the network architecture pay more attention to the temporal and spatial factors that contribute more weight to the final prediction, we use it to predict traffic speed that can better reflect the fluctuations of time and space.