Search

This study explores the dynamics of moving traffic jams at consecutive merging bottlenecks using empirical trajectory data from China's expressway. It builds on existing phase-transition point methods of Kerner's three-phase traffic theory by incorporating changepoint detection to distinguish between wide moving jams (J) and synchronized flow (S) within the expanded general congested pattern (EGP). The findings indicate that the dissolution of J in the outer lane precedes and influences the disintegration of J in the adjacent inner lane. To investigate this phenomenon, a Bayesian network (BN) was employed for causal analysis to identify key factors. Additionally, wavelet-based energy analysis was conducted on individual vehicles to examine fluctuations triggered by lane-changing maneuvers (LCMs) and their impact on the dynamics of moving jams. The results reveal that traffic flow at downstream merging bottlenecks and the lane-changing rate are crucial for J disintegration. Furthermore, the J propagated beyond the upstream merging bottleneck becomes more unstable due to the compounded effects of consecutive merging bottlenecks. The study also shows that LCMs affect the propagation of moving jams in EGP, and that the induced fluctuations may persist even after phase transitions occur. Consistent with earlier studies on German and US highways, empirical data from China show that non-regular moving jam dynamics in congested traffic, caused by LCMs and vehicle merging at adjacent bottlenecks, can be explained by complex sequences of J -> S and S -> J transitions. These sequences, which determine the empirical non-regular moving jam dynamics, exhibit qualitatively the same spatiotemporal features observed in microscopic single-vehicle data from German and US highways. Applying the phase-transition points method to empirical vehicle trajectories on China's expressways effectively reveals the spatiotemporal features of non-regular moving jam dynamics (dynamics of traffic oscillations) and traffic congestion. This research provides quantitative insights into the dynamics of traffic moving jams and the impact of LCMs, offering new strategies for mitigating congestion at complex traffic bottlenecks. • China's expressway data are studied based on Kerner's three-phase traffic theory. • Phase-transition point method combined with change-point detection identifies traffic phases. • Downstream bottleneck and lane-changing rate impact J under consecutive merging bottlenecks. • J → S and S → J transitions explain traffic congestion in China, Germany, and the USA. [ABSTRACT FROM AUTHOR]

• The probability of traffic conflicts was studied for the lane-shifting sections during the highway reconstruction and expansion activities. To accurately identify the traffic conflict in this complex road section, this study adopted the indicator pair of TTC and TDTC instead of TTC alone. • The correlations between the road and traffic factors and the vehicle collision risk were evaluated via machine learning. The influencing factors were found to be the curve radius, the standard deviation of single-vehicle speed, the cumulative angle in unit length, the vehicle type, the average speed, and the standard deviation of traffic flow speed in decreasing order of their influence on the probability of traffic conflicts. • Based on the conclusions, pertinent measures can be taken to reduce the collision risk of vehicles in the lane-shifting sections. Specific proposals are: diversion of large vehicles; restriction of the vehicle speed in the lane-shifting sections; increase of the curved section radius in line with the specifications; decrease of the cumulative turning angle per unit length i.e. the average turning curvature per unit length. Introduction: There are designated sections for lane-shifting in several highway reconstruction and expansion zones. Similar to the bottleneck sections of highways, these sections are characterized by poor pavement surface conditions, disorderly traffic flow, and high safety risk. This study examined the continuous track data of 1,297 vehicles collected using an area tracking radar. Method: The data from the lane shifting sections were analyzed in contrast with the regular section data. Further, the single-vehicle attributes, traffic flow factors, and the respective road characteristics in the lane-shifting sections were also taken into account. In addition, the Bayesian network model was established to analyze the uncertain interaction between the various other influencing factors. The K-Fold cross validation method was used to evaluate the model. Results: The results showed that the model has a high reliability. The analysis of the model revealed that the significant influencing factors in decreasing order of their influence on the traffic conflict are: the curve radius, cumulative turning angle per unit length, standard deviation of the single-vehicle speed, vehicle type, average speed, and the standard deviation of the traffic flow speed. The probability of traffic conflicts is estimated to be 44.05% when large vehicles pass through the lane- shifting section while it is 30.85% for small vehicles. The probabilities of traffic conflict are 19.95%, 34.88%, and 54.79% when the turning angles per unit length are 0.20 °/m, 0.37 °/m, and 0.63 °/m, respectively. Practical Applications: The results support the view that the highway authorities help reduce traffic risks on lane change sections by diverting large vehicles, implementing speed limits on road sections, and increasing the turning angle per unit length of vehicles. [ABSTRACT FROM AUTHOR]