Urban traffic flow prediction method based on multi-source data fusion

Abstract

The invention discloses an urban traffic flow prediction method based on multi-source data fusion. The urban traffic flow prediction method comprises the following steps: performing a fusion algorithmoriented to structured and unstructured traffic data; to be specific, designing a wavelet neural network (WNN) containing a three-layer network topology structure to process traffic volume parametersof a multi-source detector; supposing that the traffic detector performs statistics by taking 5min as a time interval, setting an acquired traffic flow sequence to be xi (i = 1, 2, 3,..., k), and using the xi as an input traffic parameter of the wavelet neural network. According to the invention, multi-azimuth detection is carried out in combination with various sensors, comprehensive and complete traffic information is obtained, structured and unstructured data are comprehensively analyzed and fused, and the accuracy and reliability of traffic flow data are improved; and in combination witha traffic flow prediction algorithm based on a deep belief network, prediction of the traffic flow is realized, so that the traffic flow prediction precision is improved and the traffic jam is relieved.

Description

technical field [0001] The invention relates to the technical field of traffic forecasting, in particular to an urban traffic flow forecasting method based on multi-source data fusion. Background technique [0002] With the continuous advancement of urbanization, there are more and more urban problems, and accurate traffic flow prediction is of great significance to solve problems such as traffic congestion and urban public safety. [0003] As an important part of the intelligent transportation system, short-term traffic flow prediction is also one of the core technologies in the development of the intelligent transportation system. Theoretical methods mainly used in short-term traffic flow prediction include: linear theoretical models based on statistics, nonlinear theoretical models and artificial intelligence models. Statistics-based linear theoretical models include historical average method, time series method, and Kalman filter method. The linear model has a simple st...

AI Technical Summary

Problems solved by technology

Due to the diversification of traffic data and the complexity of urban traffic, most of the existing traffic flow forecasting methods only use one or two data features to analyze and process the traffic flow data. to further improve

Images