
Abstract Formation control of connected and automated vehicles (CAVs) has shown great potential for improving traffic efficiency and safety. However, most existing studies focus on tracking predefined formations for fixed tasks, offering limited adaptability to changing task requirements. To address this limitation, we propose GRF-TAFC, a Grounded and Reflective Framework for Task-Adaptive Formation Choreography of CAVs. GRF-TAFC integrates context-grounded formation generation, optimization-based execution, and execution-driven reflection into a closed loop. A formation generator elicits the formation generation capabilities of pretrained MLLMs and grounds their reasoning over unstructured task inputs in real-time tool outputs, domain knowledge, and experience memory, yielding formations tailored to current task requirements while respecting environmental constraints. An execution module realizes the generated formations through polynomial-time CAV-to-target assignment and joint spatio-temporal trajectory optimization, facilitating collision avoidance and safe execution. A reflection module reasons over task context and actual execution outcomes, distilling reusable experience for iterative refinement. Real-world experiments across dynamic tasks and environments demonstrate GRF-TAFC’s effectiveness and practicality, with experimental videos available on the project website: https://dongjh20.github.io/GRF-TAFC.
Abstract Efficient freeway traffic control can alleviate congestion, improve operational efficiency, and enhance traffic safety. With the rapid development of Intelligent Transportation Systems (ITS) and connectivity technologies, Connected Vehicles (CVs), including Connected Automated Vehicles (CAVs) and Connected Human-Driven Vehicles (C-HDVs), are expected to coexist with conventional Human-Driven Vehicles (HDVs) on freeways for a considerable period. In partially connected traffic environments, CVs can provide historical and real-time Vehicle Trajectory Data (VTD) as mobile sensing sources and can also influence surrounding traffic as mobile actuators through vehicle guidance and cooperative control, thereby enabling Vehicle-Enabled Control of Traffic (VECT) for freeway flow. This paper systematically reviews recent advances in VECT under partially connected environments. The literature is organized into four categories according to the functional role of CVs and the level at which information, decisions, or control actions are generated: (1) vehicle-trajectory-enabled traffic state analysis, (2) vehicle-trajectory-enabled traffic control, (3) vehicle-guidance-enabled traffic control, and (4) collaborative vehicle–flow control. The review covers traffic state estimation, traffic state prediction, traffic anomaly detection, Variable Speed Limit (VSL) control, Ramp Metering (RM), vehicle trajectory planning, moving bottleneck control, and cross-scale coordination between macroscopic traffic regulation and microscopic vehicle guidance. Finally, key methodological challenges, deployment issues, and future research directions are discussed. This review provides a systematic framework for understanding how CVs enable freeway traffic analysis and control from sensing to actuation and coordinated vehicle–flow regulation.
Abstract Pedestrian crossing intention prediction is a critical function of autonomous vehicles, which enables safe pedestrian–vehicle interactions in complex urban environments. However, accurately predicting pedestrian intention remains challenging, as it depends on multiple factors, including pedestrian behavior, vehicle dynamics, and surrounding context. To improve prediction performance, existing approaches increasingly incorporate more modalities, but this often results in substantial computational overhead. Achieving accurate intention prediction with a limited number of modalities remains an open challenge. In this paper, we propose a vision language model-guided semantic distillation network (VLM-GSDN) for pedestrian crossing intention prediction. During training, VLM-GSDN leverages pedestrian- and scene-level semantic descriptions generated by a vision language model to align local and global visual representations with their corresponding high-level semantics, while requiring only RGB images and image-space pedestrian motion for efficient inference. To incorporate motion cues, we introduce motion-guided modulation and predictive spatial modulation, which refine pedestrian representations using pedestrian motion and predicted future displacements. A selective temporal attention module is further incorporated to integrate temporal information and emphasize informative frames. Moreover, a multi-task learning strategy is adopted in which pedestrian action-state transition prediction serves as an auxiliary task to provide additional motion supervision. Extensive experiments on the JAAD and PIE datasets demonstrate that VLM-GSDN achieves competitive performance compared with state-of-the-art methods. Moreover, it exhibits promising cross-dataset generalization and maintains robust performance over longer prediction horizons.
Abstract Bus stops vary widely in the conditions they provide to waiting passengers, yet ridership analysis rarely considers these differences alongside service performance. This study examines how stop-level amenities relate to bus ridership, conditional on service frequency and schedule deviation. Using bus stop amenity inventories, operational data, and stop-level boarding figures from Boston, USA, we have estimated a structural equation path model for a sample of continuously operated, AVL-equipped timepoint stops. The model uses observed composite scores for physical amenities at the stop level and accounts for headway and schedule deviation. The results indicate that greater provision of basic amenities is consistently linked to higher ridership, and during peak periods both amenity and service reliability measures show stronger associations. Further, spatial clustering analysis shows that poor waiting conditions are geographically concentrated rather than being randomly distributed. Taken together, the findings show how passenger waiting environments and service performance jointly influence ridership and point to opportunities for targeted investment strategies.