This paper presents the Cell Transformer (CeT), which utilizes high-definition (HD) map data to predict future traffic states at signalized intersections, thereby aiding trajectory planning for autonomous vehicles. CeT employs discretized lane segments to emulate the cell transmission model, creating a cell space to forecast vehicle counts across all segments based on historical traffic data. CeT enhances prediction accuracy by distinguishing between different vehicle types by incorporating vehicle-type attributes into vehicle-state representations through multi-head attention. In this framework, cells are modeled as nodes in a directed graph, with dynamic connections representing variations in signal phases, thereby embedding spatial relationships and signal information within dynamic graphs. Temporal embeddings derived from time attributes are integrated with these graphs to generate comprehensive spatial–temporal representations. Utilizing an encoder–decoder architecture, CeT captures dependencies and correlations from past cell states to predict future traffic conditions. Validation using real traffic data from pNEUMA demonstrates that CeT significantly outperforms baseline models in two-phase signalized intersection scenarios, achieving reductions of 11.47% in Mean Absolute Error (MAE), 13.48% in Root Mean Square Error (RMSE), and an increase of 4.36% in Accuracy (ACC). In four-phase signalized intersection scenarios, CeT shows even greater effectiveness, with improvements of 13.36% in MAE, 12.93% in RMSE, and 4.78% in ACC. These results underscore CeT’s superior predictive capabilities and highlight the contributions of its core components.

The future trajectories of surrounding agents are critical for the motion planning and control of autonomous vehicles. Thus, this study employs Transformer to develop a multi-agent trajectory prediction model named Multi-agent Trajectory Vector Transformer (MaTVT). MaTVT features a lightweight architecture, comprising a dual-level encoder formed by a low-level encoder and a high-level encoder, along with a multi-modal decoder. Once input enters MaTVT, the low-level encoder first constructs polar coordinate systems centered on target agents and then projects historical trajectories and map elements to each agent-centered coordinate system. Next, it utilizes attention mechanisms to encode motion features, agent-agent interactions, and agent-infrastructure constraints independently and fuses them into the agent encoding sequence. Considering the agent response delay, the low-level encoder extracts heterogeneous spatial-temporal features from agent encoding sequences as the local encodings for target agents. Afterward, the high-level encoder treats all agents as the nodes in a directed graph and utilizes a Graph Attention Network to convert inter-agent relationships into global encodings, which are fused with the local encodings of target agents. Finally, the multi-modal decoder translates these fusion encodings into multi-modal trajectory predictions for target agents. This study selects complex traffic scenarios from the Argoverse Motion Forecasting dataset to create a dedicated dataset for MaTVT training, validation, and testing. The test results demonstrate that MaTVT outperforms advanced benchmark methods in prediction performance, revealing its superb accuracy, efficiency, and robustness. In addition, ablation studies further explain the interpretability of the main functional components of MaTVT and their contributions to prediction performance.

Intelligent Vehicle Cyber Physical Systems (IVCPS) integrate connected vehicles, roadside infrastructure, and cloud platforms to generate digital twins of real-time urban traffic scenarios, offering significant potential for motion optimization of Intelligent Connected Vehicles (ICVs). However, existing traffic prediction methods often exhibit high latency and cannot be directly translated into motion constraints, while the cooperation framework between IVCPS and ICVs remains undefined. Therefore, this study proposes the ST-Trans Planner, a hierarchical motion optimization framework within the IVCPS-ICV cooperative environment. At the system level, the framework first constructs a dynamic directed graph with lane segments as nodes and employs a Transformer-based model integrated with the Cell Transmission Model to predict the occupancy state of lane segments. Based on these predictions, it generates a collision-free path using dynamic programming (DP) and quadratic programming (QP) and constructs a discrete spatial-temporal corridor along the path to form the initial feasible region for speed optimization. At the vehicle level, the ST-Trans Planner predicts the trajectories of surrounding traffic participants, based on which it prunes the initial solution space through probabilistic conflict detection and then applies DP and QP to obtain a smooth speed profile. The proposed framework is validated on a SUMO-CARLA co-simulation platform using real-world trajectories from the pNEUMA dataset across three urban traffic scenarios: basic segments, signalized intersections, and unsignalized intersections. Simulation experiments demonstrate that the ST-Trans Planner significantly improves both safety and efficiency for individual ICVs, achieving tighter car-following, higher passing speeds, and smoother speed profiles while satisfying real-time constraints. In cooperative optimization, the framework effectively coordinates multiple ICVs, leading to synchronized speed profiles, increased traffic throughput, and reduced idle time at intersections. Despite the increased system-level latency caused by cooperative optimization, its decoupled architecture maintains real-time responsiveness, ensuring practical deployability. Overall, this study presents a low-latency lane occupancy prediction method that directly feeds into motion constraints, a hierarchical planner that decouples path and speed optimization, and a clear division-of-labor framework for IVCPS-ICV cooperation, harnessing the advantages of both centralized and distributed control.