Jishun Ou

1. Rapid urbanization has led to severe traffic congestion in many cities, resulting in significant economic losses and adverse effects on residents’ health and quality of life. Enhancing the effic...

Understanding how congestion forms and propagates over space and time is essential for improving the operational efficiency of urban traffic systems. Recent developments in causal emergence theory indicate that the causal structures underlying dynamic models are scale-dependent. Most existing studies on traffic congestion evolution focus on a single, fixed scale, which risks overlooking clearer causal patterns at other scales and thus limiting predictive power and practical applicability. To address this, we develop a multiscale congestion evolution modeling framework grounded in causal emergence theory. Within this framework we build dynamical models at multiple spatiotemporal scales using dynamic Bayesian networks (DBNs) and quantify the causal strength of these models using effective information (EI) and singular value decomposition (SVD)-based diagnostics. Using road networks from three central Kunshan regions, we validate the proposed framework across 24 spatiotemporal scales and five demand scenarios. Across all three regions and the tested scales, we observe evidence of causal emergence in congestion evolution dynamics. When results are pooled across regions and scenarios, models built at the 10 min/150 m scale exhibit stronger and more coherent causal structure than models at other scales. These findings demonstrate that the proposed framework can identify and help build dynamical models of congestion evolution at appropriate spatiotemporal scales, thereby supporting the development of proactive traffic management and effective resilience enhancement strategies for urban transport systems.

Reliable real-time traffic state identification (TSI) provides key support for traffic management and control. Although substantial efforts have been devoted to TSI, considering the high dynamics and stochasticity of traffic flows, there remain challenges in providing reliable and consistent TSI results, especially in online network–level applications. In this study, we propose a time series clustering-based offline-online modeling framework for reliable TSI using high-resolution traffic data. Specifically, in the proposed framework, the offline module extracts representative traffic state patterns from massive historical data, which serve as the state references in the online module when performing real-time TSI with streaming information. Instead of point data, the proposed framework uses high-resolution traffic data in the form of time series, providing rich information on traffic flows and details on their short-term fluctuations and stable long-term trends. In the offline module, considering the fuzziness of traffic states, we introduce a fuzzy c-means based clustering method for offline traffic flow series clustering and traffic state pattern extraction, within which the dynamic time warping algorithm is adopted for measuring the similarity between different time series, and the optimal number of clusters is determined by a proposed critical segment–based method to reach consistent TSI in network-wide applications. In the online module, a dynamic programming–based real-time TSI approach is developed to produce reliable and smooth identification results. Extensive numerical experiments on a 20-mi-long freeway corridor in California, USA, were performed to validate the proposed framework. Results demonstrate the effectiveness of the proposed framework.

Short-term traffic flow forecasting is crucial for proactive traffic management and control. One key issue associated with the task is how to properly define and capture the temporal patterns of traffic flow. A feasible solution is to design a multi-regime strategy. In this paper, an effective approach to forecasting short-term traffic flow based on multi-regime modeling and ensemble learning is presented. First, to properly capture the different patterns of traffic flow dynamics, a regime identification model based on probabilistic modeling was developed. Each identified regime represents a specific traffic phase, and was used as the representative feature for the forecasting modeling. Second, a forecasting model built on an ensemble learning strategy was developed, which integrates the forecasts of multiple regression trees. The traffic flow data over 5-min intervals collected from four I-80 freeway segments, in California, USA, was used to evaluate the proposed approach. The experimental results show that the identified regimes are able to well explain the different traffic phases, and play an important role in forecasting. Furthermore, the developed forecasting model outperformed four typical models in terms of root mean square error (RMSE) and mean absolute percentage error (MAPE) on three traffic flow measures.

Travel time estimation plays an important role in freeway performance assessment and reliability management. Conventional estimation methods based on speed information do not consider the level of congestion during modelling, leading to unreliable estimations under congested conditions. This study presents a freeway travel time estimation method based on the general motors (GM) model, where the concept ‘virtual vehicle’ was proposed to apply macroscopic traffic flow data to the microscopic GM model. First, a GM‐based travel time estimation (GMTTE) model was developed to estimate link travel time. Three sets of parameters , and were separately used under free‐flowing, congestion and transition conditions. Next, two corridor estimation models, namely GMTTE time‐slice‐based (GMTTE‐TSB) model and GMTTE continuous‐speed (GMTTE‐CS) model, were proposed. To calibrate the three parameters, a genetic algorithm framework was developed. The proposed method was calibrated and evaluated using traffic data collected from seven freeway segments in the greater St. Louis, Missouri region. Results show that (i) the GMTTE‐CS model outperformed the GMTTE‐TSB model; (ii) the optimal parameters of the GMTTE‐CS model under free‐flowing and congested conditions are and , respectively; (iii) the GMTTE‐CS model outperformed the instantaneous and time‐slice models under congested conditions, while showing similar accuracies with the two models under free‐flowing and status‐transition conditions; and (iv) the optimal time interval was found to be 9–10 min for free‐flowing conditions and 6–7 min for congested conditions.

A common way to estimate dynamic O-D flows is to establish and solve a bilevel optimization model. Though numerous efforts have been devoted to effectively and efficiently solving the model, challenges still exist because of the interdependence of jointly solving the upper-level O-D estimation and lower-level traffic assignment problems and the nonconvexity of the model. This paper presents an alternative framework for estimating dynamic O-D flows using machine learning algorithms. The framework consists of three major modules: a learner that learns the dynamic mapping patterns describing the relationship between prior O-D flows and observed link flows, an assigner that assigns a given OD matrix to different links based on the learner, and a searcher that iteratively searches the optimal O-D solution using the assigner. A convolutional neural network is designed as the learner and trained as the assigner. Next, the algorithms to estimate a regular O-D matrix and real-time O-D flows are separately developed by using the assigner and two designed Genetic Algorithms built as the searcher. The framework was evaluated with a realistic network at the downtown area of Kunshan, China. Experimental studies show the framework can achieve satisfactory estimation performances in real time. Meanwhile, it takes raw flow ranges as the prior inputs, making it robust in the case of lacking an accurate target O-D matrix.