1 January 2024 in Journal of Advanced Transportation
During road work, lane closures significantly reduce road capacity and negatively impact traffic safety in the upstream segments. This study introduces a merge control strategy for the work zone on freeway that aims to alleviate severe congestion and improve flow efficiency in environments where manual vehicles (MVs) and connected automated vehicles (CAVs) coexist. Using a short‐term prediction model combined with a proportional‐integral‐derivative (PID) controller, this strategy dynamically adjusts merging behavior based on real‐time traffic conditions. The PID controller calculates error values as the difference between current and target states, adjusting responses through proportional, integral, and derivative terms. The predictions of the traffic state based on the density of open lanes in each segment guide the controller’s decision to initiate a “Merge” or “No Merge” guidance. When merging is deemed necessary, the controller estimates the optimal number of vehicles to merge for each segment, using the severe congestion threshold as a reference point. This approach was tested using a microscopic simulation tool on a calibrated real‐world network under mixed traffic conditions. The results indicate that the proposed strategy effectively disperses merging upstream, increases merging speeds, and maintains lane density below critical congestion levels, thus enhancing operation efficiency and safety in work zone areas.
8 August 2023 in IET Intelligent Transport Systems
A merging section on a freeway is a typical bottleneck where congestion easily occurs due to interference caused by vehicles entering the mainline from the ramp. To achieve higher speeds and/or yield to merging vehicles, drivers on the mainline often change lanes while continuously examining the circumstances ahead that are within their sight. This individual lane‐changing behaviour is somewhat myopic and could cause an imbalance in lane use, resulting in earlier onset of congestion and a more severe capacity drop than with balanced lane use. This study proposes a feedback‐based lane‐changing control strategy that balances lane flow at merge bottlenecks to mitigate capacity drop under a connected vehicle environment. The proposed Proportional‐Integral‐Derivative feedback controller examines the traffic state of the inner lanes and systematically governs the number of lane‐changing vehicles to prevent possible excessive disturbance. The proposed strategy is evaluated using microscopic simulation. For the hypothetical network, the strategy reduces the imbalance in lane flows and improves discharge flow rate as well as travel time. For the real‐world network, the strategy can also effectively relieve congestion in a possible scenario of connected automated vehicle adoption. These results indicate that the proposed strategy can improve flow efficiency for high‐volume traffic.
11 July 2024 in IET Intelligent Transport Systems
Commercial automated vehicles equipped with adaptive cruise control (ACC) systems offer multiple gap settings that determine their longitudinal behaviour. This study introduces two novel strategies—inflow control and combined control—that leverage the distinct driving behaviours associated with different gap settings in connected and automated vehicles. These strategies aim to enhance traffic efficiency in freeway lane‐drop bottlenecks, where capacity drops are common, by maintaining bottleneck occupancy at the target level using a proportional‐integral‐derivative controller. Simulation experiments were conducted using VISSIM to validate the proposed strategies. The results from a hypothetical lane‐drop bottleneck indicate that the proposed strategies enhanced both efficiency and safety across all simulated demand levels, with the combined control outperforming inflow control by redistributing the relative positions of vehicles before the mandatory lane changes using a new gap setting. Moreover, the proposed strategies were effective under all the simulated market penetration rates (MPRs), where better performances were demonstrated at higher MPRs. An evaluation of a calibrated real‐world network further demonstrated the potential of recommending gap settings to drivers of ACC‐equipped vehicles using variable message signs to enhance freeway efficiency in the near future.
8 April 2026 in Transportation Research Part B: Methodological
3 February 2026 in SSRN Electronic Journal
Traffic flow heterogeneity, stemming from diverse behaviors of its agents, presents a fundamental challenge in developing accurate predictive and prescriptive models that are at the heart of understanding traffic. This paper adopts a data decomposition framework to address the limitations of global data aggregation methods that often obscure individual agent behaviors in traffic systems. We propose a decomposition framework that represents each agent's data as a sum of population-level common component, and agent-specific specific personalized component. The common component is learned as a shared temporal subspace across agents, while personalized components are extracted in an orthogonal complement, yielding an interpretable separation between common and personalized behavior. Through stylized experiments we show that pooled learning can collapse under heterogeneity, misleading traffic-level interpretation. Whereas the proposed decomposition recovers agent-level behavior, enhancing interpretability. We further apply the method to empirical driving data to perform behavioral analysis, personalization, and behavioral clustering across different agents. The framework provides a bottom-up bridge between traffic agent heterogeneity and traffic-level analysis.
5 March 2026 in SSRN Electronic Journal
As connected automated vehicles become integrated into modern transportation systems, cybersecurity vulnerabilities increasingly threaten traffic operations and safety. Falsified lane-changing (LC) behavior, induced by cyberattacks such as false data injection, can disrupt proper decision-making and degrade system performance. This study proposes a framework for detecting abnormal LC decisions under cyberattack. The framework comprises two modules: (i) LC decision prediction and (ii) detection of falsified behavior. For LC prediction, this study proposes a physics-guided neural network (PGNN) framework. The PGNN framework integrates a game theory-based LC decision model with a long short-term memory (LSTM) network, complementing physics-based interpretability with data-driven learning capability. Falsified behavior is then detected by monitoring cumulative deviations between predicted LC probabilities and the observed behavior. Evaluation using real-world trajectory data shows that the proposed method achieves high recall and strong overall accuracy under the evaluated attack settings, outperforming existing baselines and a standalone LSTM-based detection.