traffic microsimulation and vehicle-level modeling

To address insufficient traffic simulation fidelity and ambiguous requirement specifications in driving simulators, this paper proposes a systematic traffic simulation requirement analysis method based on sub-goal decomposition. The experimental objective is hierarchically decomposed into verifiable sub-goals—including microscopic traffic modeling, agent behavioral modeling, and visual rendering—thereby establishing a structured, traceable mapping from research objectives to simulation configuration. This method establishes, for the first time, an explicit linkage between traffic simulation design and underlying experimental goals, significantly enhancing simulation fidelity, experimental validity, and participant immersion. Empirical evaluation demonstrates that the proposed framework supports high-fidelity development and human–autonomy interaction testing of autonomous driving systems.

This study addresses the limitation of conventional microsimulation in faithfully reproducing transient traffic wave dynamics. We propose a data-driven co-simulation framework integrating CARLA with high-fidelity trajectory data from the I-24 MOTION dataset. Our method introduces a boundary-condition-driven mechanism—leveraging ghost cells, autonomous vehicle generation, and configurable vehicle dynamics models—to reconstruct spatiotemporal traffic states end-to-end. To our knowledge, this is the first implementation of measurement-based, boundary-driven simulation within CARLA, overcoming the constraints of localized car-following models and enabling microscale emergence of macroscopic traffic phenomena. Experiments successfully replicate the formation, propagation, and dissipation of traffic waves under both high- and low-density conditions, achieving significantly improved spatiotemporal fidelity. The framework provides a high-fidelity simulation environment for evaluating traffic control strategies and validating autonomous vehicle perception systems.

Existing traffic simulators struggle to generate large-scale, high-fidelity urban scenarios with diverse driving styles, hindering robust evaluation of autonomous driving systems. To address this, we propose the first controllable traffic simulation framework based on diffusion models. Our method unifies multi-source traffic data into a coherent scene representation and introduces a conditional guidance mechanism to explicitly control driving styles—including aggressive, conservative, and stochastic behaviors. This approach overcomes the dual limitations of conventional rule-based and data-driven simulators in behavioral diversity and scene generalizability, enabling real-time, high-fidelity simulation over city-scale road networks with thousands of vehicles. Experiments demonstrate that our platform supports systematic manipulation of key variables—such as traffic density and style composition—thereby significantly enhancing the effectiveness and interpretability of autonomous driving algorithm evaluation, particularly in edge-case scenarios.

Existing microscopic traffic simulation relies on single real-world datasets, resulting in insufficient scenario diversity and limiting the training and evaluation efficacy of autonomous driving algorithms. Method: We propose an infinite-diversity traffic scenario generation framework tailored for autonomous driving, introducing a novel two-stage paradigm that synergistically integrates large language models (LLMs) and vision-language models (VLMs): semantic scene planning driven by LLMs in Stage I, followed by high-fidelity trajectory synthesis guided by VLMs in Stage II. We further design DriveGen-CS—a fine-tuning-free method that leverages algorithmic failure feedback to automatically trigger long-tail and edge-case scenario generation. The framework incorporates retrieval-augmented generation (RAG), diffusion-based planners, and customized trajectory modeling. Results: Experiments demonstrate significantly superior scenario diversity over state-of-the-art methods, improved downstream driving policy performance, and a 37% increase in edge-case detection rate.

Existing generative vehicle trajectory models are predominantly evaluated using reconstruction error, lacking online validation of traffic compliance and safety within real-time microscopic simulation—and overlooking critical engineering issues such as red-light running and illegal parking. This paper proposes the first online evaluation framework for generated trajectories at signalized intersections: it integrates SUMO as a closed-loop microscopic simulator to enable online assessment under unseen traffic conditions; introduces traffic-rule-aware metrics—including red-light violation rate and illegal parking frequency; and establishes a quantitative analysis toolchain. Experiments reveal that state-of-the-art models, despite low reconstruction error, frequently generate unsafe and noncompliant trajectories. The proposed metrics significantly enhance detection of safety-critical failures, delivering a reproducible, interpretable, and engineering-deployable evaluation paradigm for trustworthy traffic AI. (149 words)

This work addresses the challenge that traditional traffic microsimulation tools, due to oversimplified behavioral models, struggle to generate physically consistent long-horizon trajectories—particularly in critical areas such as intersections. To overcome this limitation, the study introduces, for the first time, a world model paradigm into traffic simulation by proposing an agent-centric generative world model that integrates a Transformer architecture with intersection geometry awareness to jointly model agent interactions and dynamic environmental context. A closed-loop validation framework is implemented within SUMO, demonstrating that the model can stably produce high-fidelity trajectories using only limited training data. Experimental results over 4000-second simulations show significant improvements over baseline methods in both traffic engineering metrics and aggregate performance indicators, with a more than tenfold reduction in KL divergence, thereby validating the approach’s breakthroughs in realism, consistency, and data efficiency.

Traditional rule-based microscopic traffic simulation struggles to accurately reproduce real-world traffic conflict dynamics, limiting its ability to predict collision frequencies. This study proposes a novel approach that integrates machine learning–based driving behavior models with microscopic simulation to generate high-fidelity conflict scenarios. By combining a two-dimensional time-to-collision (TTC) metric with extreme value theory, the method enables precise estimation of intersection collision frequencies. Notably, this work demonstrates for the first time that a generic machine learning driving behavior model—without site-specific calibration—can significantly enhance the accuracy of surrogate safety assessments based on traffic conflicts, whereas conventional rule-based models fail to capture actual collision trends effectively.

This work addresses the limitations of existing microscopic traffic simulation models in accurately capturing heterogeneous vehicle interactions at signalized intersections and the myopic, unstable behavior of learning-based trajectory predictors under closed-loop execution. To overcome these challenges, the authors propose Enactor—a vehicle-centric, generative closed-loop microscopic simulation framework. Enactor encodes dynamic agents and lane markings in polar coordinates, employs a spatiotemporally decoupled attention mechanism within a Transformer architecture to predict motion distributions, and leverages a closed-loop curriculum training strategy. The method achieves, for the first time, long-term stable generative simulation at intersections: in 4000-second closed-loop tests, it reduces the KL divergence of speed and travel time distributions by over fivefold compared to baselines, decreases red-light violations by more than an order of magnitude, and significantly outperforms constant-velocity baselines in multi-step prediction accuracy.

Existing traffic simulation tools struggle to accurately capture the complex interactions between human-driven and autonomous vehicles in mixed traffic environments and lack a systematic synthesis of relevant AI methodologies. This work proposes the first unified taxonomy that encompasses three categories of AI approaches: agent-level behavioral modeling, environment-level simulation, and integration of cognitive and physical information, thereby bridging the research gap between transportation engineering and computer science. By consolidating mainstream simulation platforms, datasets, and evaluation metrics, the study systematically analyzes the limitations of current tools, clarifies the evolutionary trajectory of AI methods in this domain, and puts forward a standardized evaluation protocol along with promising future research directions to advance high-fidelity mixed traffic simulation.

This study addresses the challenge of holistically evaluating urban traffic control policies, where direct effects—such as changes in traffic flow and emissions—are intricately intertwined with indirect effects, including behavioral responses and shifts in economic accessibility. To this end, the authors propose a multilayer urban mobility simulation framework that integrates a physical layer (modeling traffic dynamics and emissions) with a social layer (capturing user behavioral responses). The framework leverages real-world data to instantiate scenarios, encode policy parameters, and formalize behavioral assumptions, thereby enabling systematic comparison and forward-looking assessment of diverse “what-if” policy scenarios. Applied to vehicle restriction policies, the approach effectively uncovers the interactive mechanisms between policy design and user feedback, offering actionable insights for developing more anticipatory and coordinated transportation policies.