Time-of-Day Breakpoints Identification And Traffic Network Partition Methods For Traffic Control

The volatility and uncertainty of traffic flow is one of the main causes of traffic control complexity. Such volatility in the traffic flow is reflected in both time and space. This paper studies the Time-of-Day breakpoints identification and sub-area division methods. The sub-area division can be divided into two types according to the control area size:the coordinated intersections division used in arterial coordination control and sub-area division used in regional coordination control.As for Time-of-Day breakpoints identification, the effects of the clustering data composition and interval were analyzed at first. Then the improved NJW spectral clustering algorithm is used for Time-of-Day breakpoints identification in an intersection and multiple intersections case. The obtained result was compared with that of K-means algorithm by simulation. Considering the use of multi-day traffic flow data, identification procedure was renewed. Principal component analysis was used for data dimensionality reduction and Self-Tuning spectral clustering algorithm was applied for Time-of-Day breakpoints identification. The clustering index and simulation results both demonstrate the effectiveness and validity of this method. Then taking into account the timing characteristics of the traffic flow, it was found that the use of sequential cluster analysis of on the optimal period was better than that of the K-means algorithm.In terms of coordinated intersections division, the calculation model of Coordinatability Factor was fully analyzed which was used to describe the intersections relevance in Synchro software. Then in a case study, the procedures and methods of coordinated intersections division using Synchro software was analyzed. The simulation results show Synchro software can not directly obtain the optimal division, and division coordinated intersection has a significant impact on the coordination control effect. Then based on the model of maximum number through the intersection, the establishment of criteria for the classification coordination intersection and the method is simple and effective.In terms of the sub-area division, firstly simulation environment was established based on microsimulation software VISSIM to analyze how the changes of intersection cycle, green ratio and saturation level affect optimal signal timing and control effectiveness of adjacent ones, which contribute to further understand the relevance of the intersections. Then an improved algorithm based on community detecting methods in complex networks was applied to weighted networks and sub-area division method was proposed in a large scale transport network.