Traffic prediction method based on comprehensive spatial-temporal characteristics

• Main Authors: MA LUJUAN, ZHOU JIANLIN, DENG XIAOPING
• Format: Patent
• Language: Chinese; English
• Published: 10.05.2024
• Subjects: PHYSICS; SIGNALLING; TRAFFIC CONTROL SYSTEMS

The invention discloses a traffic prediction method based on comprehensive spatio-temporal characteristics, which comprises the following steps: extracting spatio-temporal dimension characteristics of a predicted road section based on a traffic network, and inputting the spatio-temporal dimension characteristics as an original image signal matrix into a spatio-temporal block for processing to obtain a final input quantity, the spatio-temporal block comprising a spatio-temporal attention module and a spatio-temporal convolution module; combining the final input quantity with a result obtained by carrying out residual error convolution operation on the original image signal matrix, and processing to obtain a final output quantity; and inputting the final output quantity into the LSTM network model, and inputting an output result of the model into a full connection layer to obtain a final prediction result. According to the invention, the traffic management efficiency can be improved, and the possibility of cong