A Two-step Prediction Method of Dynamic Turning Ratio at an Intersection Based on Double Bayesian

Abstract

The invention discloses a crossroad dynamic turning proportion two-step prediction method based on double Bayes. According to the crossroad dynamic turning proportion two-step prediction method, a first Bayes combination method is designed by means of link flow detected by entries and exits at a crossroad and with combination of history flow data to predicate entry and exist flow in the next time period, an improved Kalman filtering algorithm and an improved counterpropagation neural network algorithm are designed based on the entry and exist flow in the next time period to predicate dynamic turning proportion of the next first time period and the next second time period, a second Bayes combination method is designed under the conditions that predication errors are corrected through history turning proportion data to calibrate and update weight dynamically, and dynamic turning proportion one-step and two-step prediction values of double Bayes combination methods can be obtained. By means of the existing method, dynamic turning proportion one-step value can be obtained only, and the existing method has advantages and disadvantages in precise and efficiency. The crossroad dynamic turning proportion two-step prediction method comprehensively has the advantages of all methods, avoids local overhigh deviation, is high in precision, and can obtain one-step and two-step prediction values simultaneously and provide basis supporting for an intelligent traffic control system.

Description

technical field [0001] The invention belongs to the technical field of intelligent traffic control, and in particular relates to a two-step prediction method for the dynamic turning ratio of an intersection based on double Bayesian, which is used for the development of an intelligent signal control system at the intersection. Background technique [0002] As an important node of the urban road network, intersections have the characteristics of non-linear and time-varying traffic flow in each direction. Scientific and reasonable intersection signal control and traffic organization schemes should be based on accurate and real-time traffic volumes, while dynamic steering flows, especially in the future The forecast value of dynamic steering flow in time period is the basic data of intersection intelligent signal control. Under the conditions of existing traffic flow detection technology, it is easy to obtain the traffic flow of each lane upstream of the entrance road and downst...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is that the existing technology cannot detect the dynamic steering ratio of the intersection in real time; some methods can only estimate the steering ratio in the current period in real time, but cannot predict; among the only two methods with prediction functions, Carl The Mann filtering method can only predict single-step results, while the BP neural network method has slow calculation speed and is easy to fall into local optimal results

Images