A Dynamic Regional Partitioning Method for Active Traffic Control

Abstract

In order to identify the scope of active traffic control regions and improve the effect of active traffic control, this paper proposes a dynamic partitioning method of area boundaries based on benchmark intersections, taking into account the saturation, homogeneity, and correlation of intersections in the region. First, a boundary indicator correlation model was established. Next, benchmark intersections were selected based on evaluation indicators, such as traffic speed and queue length. Then, the boundary of the control region is initially defined based on the selected reference intersection, through a combination of the improved Newman algorithm. Subsequently, a spectral clustering algorithm is used to obtain the boundaries of the optimal active control subregions. Finally, a city road network is used as the study object for analysis and verification under the premise of implementing active traffic control. The results show that compared with the intersection clustering algorithm method and the boundary control subdivision method, the control effect indicators, such as the average delay and the average number of stops, have a great optimization improvement. Thus, the proposed method of regional borders combines the actual traffic flow characteristics efficiently to make a more accurate real-time dynamic division of the road network sub-areas.

A Study of Adjacent Intersection Correlation Based on Temporal Graph Attention Network

“…In terms of the traffic state, in the existing research, it is mainly transformed into a clustering problem through the extraction of the feature vectors of the traffic state from traffic data, such as speed, flow rate, time occupancy, etc., and then artificial intelligence algorithms are used to solve the problem, such as the spectral clustering algorithm [3], k-means algorithm [4], DBSCAN algorithm [5], etc. The shortcomings of the existing research are as follows: (1) The object of many studies is the traffic state of road sections, not the traffic state of intersections.…”