Urban Road Traffic Flow Prediction Based On Spatial And Temporal Characteristics

In recent years,the rapid growth of China's GDP and population urbanization has prompted the Chinese government to increase investment in the transportation industry,and the scale of the urban road network has been developed accordingly.However,due to the influence of external environmental factors such as urban terrain and large buildings,the urban road network cannot be expanded at will.At this stage,urban residents' motor vehicle ownership is growing rapidly,resulting in the inconsistency between the infrastructure construction speed of urban road networks and the growth rate of motor vehicle ownership,resulting in serious traffic congestion.Real-time and accurate urban road traffic information prediction has become the key to alleviate traffic congestion.Most of the traditional traffic flow prediction methods only consider the single characteristics of urban road traffic flow in time or space in the modeling process,and lack the in-depth analysis of the temporal and spatial characteristics of urban road traffic flow.As a result,the existing traffic flow prediction models are not accurate enough to timely update the parameters and other problems.In view of the above problems,this dissertation takes the traffic volume on urban roads as the research object,and starts with the spatio-temporal characteristics of urban road traffic flow.It analyzes the urban road network structure and the temporal and spatial characteristics of urban road traffic flow in detail.The influence of the traffic flow of adjacent roads was added when constructing the prediction model of urban road traffic flow.The urban road traffic flow prediction method which considered the historical traffic data of the target road section and the spatial traffic flow data of adjacent road sections was designed,which improves the accuracy of urban road traffic flow prediction.The core content of this dissertation is as follows:Firstly,the temporal and spatial characteristics of urban road traffic flow are deeply analyzed,and the existing urban road traffic flow prediction methods are summarized.Secondly,the structural characteristics of urban road network and the temporal and spatial characteristics of traffic flow on the road are analyzed.Finally,two prediction models based on the analysis of the spatial and temporal characteristics of urban road traffic flow,LSTM-RF model and GCN+LSTM model,are constructed.The first LSTM-RF model uses the good linear fitting ability of LSTM and the strong generalization ability of RF to deeply analyze the traffic flow data of the target road section and the traffic flow data of the adjacent road section spatially,and then using the idea of decision-level data fusion to effectively integrate the prediction results of the two models to achieve urban road traffic flow prediction.The second model,GCN+LSTM model,first uses the weight-sharing feature of the GCN network to extract the spatial characteristics of the traffic flow of the target road.Secondly,it uses the memory characteristics of the LSTM network to extract the periodic information of the target road,and then through the GCN and LSTM Appropriate combination realizes the fusion of spatio-temporal features of urban road traffic flow.Finally,the actual data verifies the prediction performance of the prediction method proposed in this dissertation.