The DOI matches its Crossref record; 55 of the 64 references it cites were found in a registry or an index.
SceneGen: Learning to Generate Realistic Traffic Scenes
Open accessAbstract
We consider the problem of generating realistic traffic scenes automatically. Existing methods typically insert actors into the scene according to a set of hand-crafted heuristics and are limited in their ability to model the true complexity and diversity of real traffic scenes, thus inducing a content gap between synthesized traffic scenes versus real ones. As a result, existing simulators lack the fidelity necessary to train and test self-driving vehicles. To address this limitation, we present SceneGen—a neural autoregressive model of traffic scenes that eschews the need for rules and heuristics. In particular, given the ego-vehicle state and a high definition map of surrounding area, SceneGen inserts actors of various classes into the scene and synthesizes their sizes, orientations, and velocities. We demonstrate on two large-scale datasets SceneGen’s ability to faithfully model distributions of real traffic scenes. Moreover, we show that SceneGen coupled with sensor simulation can be used to train perception models that generalize to the real world.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“Generating large synthetic datasets has been a problem of great interest in other domains as well [71-77], and has also been used as a tool for benchmarking motion planning [78].”
From MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations (2023) · cites this paper as [77]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 1×
“The basis of our model is the ConvLSTM architecture [42]—an extension of the classic LSTM architecture [22] to spatial data—and the input to our model at the i-th generation step is a bird’s eye view multi-channel image encoding the SDV a0, the HD map m, and the actors generated so far {a1, .”
MethodAdam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
Book Review: Pattern Recognition and Machine Learning
Nasrabadi, N. M. (2007) · Journal of Electronic Imaging · doi:10.1117/1.2819119 · Cited by 22,066 · cited 1×
“Each actor’s bounding box is sampled from a Gaussian KDE fitted to the training dataset [5] and velocities are set to satisfy speed limits and a time gap between successive actors on the lane graph.”
MethodGANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., & Hochreiter, S. (2017) · arXiv · doi:10.48550/arxiv.1706.08500 · Cited by 8,365
Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting
Shi, X., Chen, Z., Wang, H., Yeung, D.-Y., Wong, W.-K., & Woo, W.-C. (2015) · arXiv · doi:10.48550/arxiv.1506.04214 · Cited by 7,709
Improved Techniques for Training GANs
Salimans, T., Goodfellow, I., Zaremba, W., Cheung, V., Radford, A., & Chen, X. (2016) · arXiv · doi:10.48550/arxiv.1606.03498 · Cited by 6,965 · cited 1×
“This is akin to the popular IS [41], FID [21], and KID [4] metrics for evaluating generative models, except we use a feature extractor trained on traffic scenes.”
MethodWaveNet: a generative model for raw audio
Oord, A. V. D., Dieleman, S., Zen, H., Simonyan, K., Vinyals, O., Graves, A., Kalchbrenner, N., Senior, A., & Kavukcuoglu, K. (2016) · arXiv · doi:10.48550/arxiv.1609.03499 · Cited by 4,911 · cited 1×
“Recently, neural autoregressive models have found tremendous success in modeling a variety of data, including handwriting [18], images [49], audio [48], text [39], sketches [20], graphs [29], 3D meshes [33], indoor scenes [50] and image scene layouts [25].”
BackgroundCongested traffic states in empirical observations and microscopic simulations
Treiber, M., Hennecke, A., & Helbing, D. (2000) · Physical Review E · doi:10.1103/physreve.62.1805 · Cited by 4,762 · cited 1×
“In contrast, microscopic models simulate the behavior of each individual vehicle over time by assuming a car-following model [36, 6, 34, 13, 17, 1, 44].”
BackgroundOn kinematic waves II. A theory of traffic flow on long crowded roads
Lighthill, S. J., & Whitham, G. B. (1955) · Proceedings of the Royal Society of London A Mathematical and Physical Sciences · doi:10.1098/rspa.1955.0089 · Cited by 4,606
Microscopic Traffic Simulation using SUMO
López, P. Á., Wießner, E., Behrisch, M., Bieker-Walz, L., Erdmann, J., Flötteröd, Y.-P., Hilbrich, R., Lücken, L., Rummel, J., & Wagner, P. (2018) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc.2018.8569938 · Cited by 3,898 · cited 7×
“Microscopic traffic models have been included in popular software packages such as SUMO [31], CORSIM [35], VISSIM [11], and MITSIM [55].”
BackgroundMethodInfluential
Research works printed but not linked: 3
- gerlough, 1955, Simulation of Freeway Traffic on a General-Purpose Discrete Variable Computer
- uria, 2013, NADE: the real-valued neural autoregressive density-estimator, CoRR
- taheri, 1990, An investigation and design of slip control braking systems integrated with four wheel steering
Cited by
Sorted by citations
End-to-End Autonomous Driving: Challenges and Frontiers
Chen, L., Wu, P., Chitta, K., Jaeger, B., Geiger, A., & Li, H. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3435937 · Cited by 843
“They can then generate novel scenarios that resemble the real world but were not included in the original data [96], [97], [98], [99].”
BackgroundGenerative artificial intelligence: a systematic review and applications
Sengar, S. S., Hasan, A. B., Kumar, S., & Carroll, F. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-20016-1 · Cited by 763
“The model outperforms the previous state-of-art techniques such as SeneGen [127] in generating the synthetic traffic scenarios.”
3D Object Detection for Autonomous Driving: A Comprehensive Survey
Mao, J., Shi, S., Wang, X., & Li, H. (2023) · International Journal of Computer Vision · doi:10.1007/s11263-023-01790-1 · Cited by 529
“There are also some works simulating the traffic flows (Tan et al., 2021; Suo et al., 2021) or testing the safety of vehicles by simulation (Wong et al.”
BackgroundV2X-Sim: Multi-Agent Collaborative Perception Dataset and Benchmark for Autonomous Driving
Li, Y., Ma, D., An, Z., Wang, Z., Zhong, Y., Chen, S., & Feng, C. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3192802 · Cited by 381
A Survey on Safety-Critical Driving Scenario Generation—A Methodological Perspective
Ding, W., Xu, C., Arief, M., Lin, H., Li, B., & Zhao, D. X. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3259322 · Cited by 320
“introduced into the generation in [36] and [37].”
BackgroundExplainability of Deep Vision-Based Autonomous Driving Systems: Review and Challenges
Zablocki, É., Ben-Younes, H., Pérez, P., & Cord, M. (2022) · International Journal of Computer Vision · doi:10.1007/s11263-022-01657-x · Cited by 227
“, 2017), by inserting virtual objects in real scenes (Alhaija et al., 2018; Tan et al., 2021), by augmenting existing images with dynamic objects extracted from other scenes (Chen et al.”
BackgroundXu, H., Omitaomu, O. A., Sabri, S., Sisi Zlatanova, B., Li, X., & Song, Y. (2024) · Urban Informatics · doi:10.1007/s44212-024-00060-w · Cited by 253
Multi-Sensor Fusion Technology for 3D Object Detection in Autonomous Driving: A Review
Wang, X., Li, K., & Chehri, A. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3317372 · Cited by 179
“It is worth mentioning that Tan et al. [104] proposed a method to automatically generate realistic traffic scenes with the flexibility to insert various classes of objects in the desired scenes.”
MethodTrafficGen: Learning to Generate Diverse and Realistic Traffic Scenarios
Feng, L., Li, Q., Peng, Z., Tan, S., & Zhou, B. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10160296 · Cited by 149
“Compare to SceneGen, our method does not need post-processing and can augment existing traffic scenarios.”
BackgroundMethodInfluentialSelf-citationSynthetic Datasets for Autonomous Driving: A Survey
Song, Z., He, Z., Li, X., Ma, Q., Ming, R., Mao, Z., Pei, H., Peng, L., Hu, J., Yao, D., et al. (2023) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2023.3331024 · Cited by 120
Choose Your Simulator Wisely: A Review on Open-Source Simulators for Autonomous Driving
Li, Y., Yuan, W., Zhang, S., Yan, W., Shen, Q., Wang, C., & Yang, M. Z. (2024) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2024.3374044 · Cited by 95
“These techniques prioritize comprehensively depicting the traffic environment, encompassing traffic elements such as the road network, traffic signals, and dynamic objects [108], [148].”
BackgroundData-Driven Traffic Simulation: A Comprehensive Review
Chen, D., Zhu, M., Yang, H., Wang, X., & Wang, Y. (2024) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2024.3367919 · Cited by 86
“Tan, et al. [95] presented a neural autoregressive model named SceneGen, which inserts actors (vehicles) of di-verseclassesintothesceneandgeneratestheirsizes,orientations,andvelocities.”
BackgroundMethodV2XP-ASG: Generating Adversarial Scenes for Vehicle-to-Everything Perception
Xiang, H., Xu, R., Xia, X., Zheng, Z., Zhou, B., & Ma, J. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10161384 · Cited by 39
RealGen: Retrieval Augmented Generation for Controllable Traffic Scenarios
Ding, W., Cao, Y., Zhao, D. X., Xiao, C., & Pavone, M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73033-7_6 · Cited by 48
Deep Learning for Traffic Scene Understanding: A Review
Dolatyabi, P., Regan, J., & Khodayar, M. (2025) · IEEE Access · doi:10.1109/access.2025.3529289 · Cited by 35
Yang, L., Yuan, J., Zhao, X., Fang, S., He, Z., Zhan, J., Hu, Z., & Li, X. (2023) · Journal of Intelligent and Connected Vehicles · doi:10.26599/jicv.2023.9210023 · Cited by 28
Overview of image-based 3D reconstruction technology
Niu, Y., Liu, L., Huang, F., Huang, S., & Chen, S. (2024) · Journal of the European Optical Society Rapid Publications · doi:10.1051/jeos/2024018 · Cited by 33
1001 Ways of Scenario Generation for Testing of Self-driving Cars: A Survey
Schütt, B., Ransiek, J., Braun, T., & Sax, E. (2023) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv55152.2023.10186735 · Cited by 32
“Both [19] and [20] address the problem of generating realistic traffic scenes and solely operate on layer 4.”
BackgroundSelf-Supervised Real-to-Sim Scene Generation
Prakash, A., Debnath, S., Lafleche, J.-F., Cameracci, E., State, G., Birchfield, S. T., & Law, M. T. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01574 · Cited by 23
“SceneGen [52] learns to generate traffic scenes by modeling the attributes of all objects.”
BackgroundMethodGINA-3D: Learning to Generate Implicit Neural Assets in the Wild
Shen, B., Yan, X., Qi, C. R., Najibi, M., Deng, B., Guibas, L. J., Zhou, Y., & Anguelov, D. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00476 · Cited by 23
“Many recent works explored different aspects of data-driven simulation, including image synthesis [114–117], assets modeling [47,48,118–121], scene generation [49,122,123], and scene rendering [1,124–126].”
BackgroundA Survey of Scenario Generation for Automated Vehicle Testing and Validation
Wang, Z., Ma, J., & Lai, E. M.-K. (2024) · Future Internet · doi:10.3390/fi16120480 · Cited by 22
Solving Motion Planning Tasks with a Scalable Generative Model
Hu, Y., Chai, S., Yang, Z., Qian, J., Li, K., Shao, W., Zhang, H., Xu, W., & Liu, Q. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72946-1_22 · Cited by 26
LidarDM: Generative LiDAR Simulation in a Generated World
Zyrianov, V., Che, H., Liu, Z., & Wang, S. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11128001 · Cited by 56
“More recently, they have shown great promise in robotics by generating realistic scenarios and sensory data for training and validating embodied agents, significantly reducing the cost of real-world experiments [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23].”
Self-citationDriveSceneGen: Generating Diverse and Realistic Driving Scenarios From Scratch
Sun, S., Gu, Z., Sun, T., Sun, J., Yuan, C.-M., Han, Y., Li, D., & Ang, M. H. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3416792 · Cited by 26
“Works including [16]–[18] focused on sampling agent states from existing scenarios, then erasing and respawning the agents based on agent state distribution.”
BackgroundSceneControl: Diffusion for Controllable Traffic Scene Generation
Lu, J., Wong, K., Zhang, W., Suo, S., & Urtasun, R. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10610324 · Cited by 19
“Most also focus on generating scenes unconditionally [2], [37], [29], [8].”
BackgroundMethodInfluentialSelf-citationLong Short Term Memory Based Traffic Prediction Using Multi-Source Data
Leinonen, M., Al-Tachmeesschi, A., Turkmen, B., Atashi, N., & Ruotsalainen, L. (2024) · International Journal of Intelligent Transportation Systems Research · doi:10.1007/s13177-024-00451-y · Cited by 17
“On the other hand, microscopic traffic simulations aim to simulate the behaviour of each individual vehicle [11].”
Liu, X., Huang, H., Bian, J., Zhou, R., Wei, Z., & Zhou, H. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.111995 · Cited by 14
RSG-GCN: Predicting Semantic Relationships in Urban Traffic Scene With Map Geometric Prior
Tian, Y., Carballo, A., Li, R., & Takeda, K. (2023) · IEEE Open Journal of Intelligent Transportation Systems · doi:10.1109/ojits.2023.3260624 · Cited by 14
“…parameterize actor node α i ∈ A with an eight-dimensional vector containing its class label c i ∈ C , its position according to the ego-centric bird’s-eye-view coordinates ( x i , y i ) ∈ R 2 , its bounding box b i ∈ B , and its velocity ( v ix , v iy ) ∈ R 2 , based on the method proposed in [44].”
MethodInfluentialGOOSE: Goal-Conditioned Reinforcement Learning for Safety-Critical Scenario Generation
Ransiek, J., Plaum, J., Langner, J., et al. (2024) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc58415.2024.10919498 · Cited by 11
“[22] propose a neural autoregressive model to insert actors of different classes into traffic scenes.”
Text-to-Drive: Diverse Driving Behavior Synthesis via Large Language Models
Nguyen, P., Wang, T.-H. J., Hong, Z.-W., Karaman, S., & Rus, D. L. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10801754 · Cited by 10
“Then, to evaluate the alignment of driving policies with desired behaviors, we introduce the primary reward function (subsection III-D), which, together with our auxiliary function, provides reward guidance to the driving policy (subsection III-E).”
Zhang, G., Huang, H., Zhou, R., Li, S., & Bian, J. (2025) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2025.3576236 · Cited by 12
“Tan et al. modeled trajectory features using LSTM and compared the effectiveness of GAN and VAE for synthesizing trajectories [41].”
Lift3D: Synthesize 3D Training Data by Lifting 2D GAN to 3D Generative Radiance Field
Li, L., Lian, Q., Wang, L., Ma, N., & Chen, Y.-C. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00040 · Cited by 12
“Based on the proposed framework, we can enhance original datasets both in quantity and diversity.”
SYGNet: A SVD-YOLO based GhostNet for Real-time Driving Scene Parsing
Wang, H., Zhu, B., Li, Y., et al. (2022) · Proceedings - International Conference on Image Processing · doi:10.1109/icip46576.2022.9897534 · Cited by 9
“Autonomous driving field consists of scenario generation [1,2], scene recognition [3,4], navigation [5], simulation [6], scene parsing [7–9], prediction [10, 11] and motion estimation [12, 13].”
BackgroundSelf-citationSemantically Adversarial Scene Generation With Explicit Knowledge Guidance
Ding, W., Lin, H., Li, B., & Zhao, D. X. (2025) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3510515 · Cited by 16
Quantile-based scenario generation for automated vehicle safety evaluation
Zhou, H., Ma, C., Ma, K., et al. (2025) · Accident Analysis & Prevention · doi:10.1016/j.aap.2025.108043 · Cited by 9
Sim-on-Wheels: Physical World in the Loop Simulation for Self-Driving
Shen, Y., Chandaka, B., Lin, Z.-H., Zhai, A. J., Cui, H., Forsyth, D. A., & Wang, S. (2023) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2023.3325689 · Cited by 9
“data-driven simulation in [11], [12], [14], [15]; dynamic models in [33], [34], [35] ; environments in [13], [36], [37], [38]).”
Self-citationDreamCar: Leveraging Car-Specific Prior for In-the-Wild 3D Car Reconstruction
Du, X., Sun, H., Lu, M., Zhu, T., & Yu, X. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3523231 · Cited by 7
“They must deal with infrequent dangerous situations, but collecting such dangerous data is challenging in the real-world setting [1].”
BackgroundHAD-Gen: Human-like and diverse driving behavior modeling for controllable scenario generation
Wang, C., Kong, L., Tamborski, M., & Albrecht, S. V. (2025) · Accident Analysis & Prevention · doi:10.1016/j.aap.2025.108270 · Cited by 9
“An example is the SceneGen proposed by Tan et al. [33], which aims to generate realistic traffic scenes by learning from large-scale real-world datasets.”
LLM-Powered Synthetic Environments for Self-Driving Scenarios
Adekanye, O. A. M. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i21.30540 · Cited by 7
EYOLOX: An Efficient One-Stage Object Detection Network Based on YOLOX
Tang, R., Sun, H., Liu, D., Xu, H. F., Qi, M., & Kong, J. (2023) · Applied Sciences · doi:10.3390/app13031506 · Cited by 7
MJTG: A Multi-Vehicle Joint Trajectory Generator for Complex and Rare Scenarios
Tian, Y., Zheng, W., Shao, Y., et al. (2025) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2025.3567636 · Cited by 6
“Tan, et al. generated realistic scenarios by inputting vehicle states and map information into a Long Short Term Memory network (LSTM) [43].”
Realistic Corner Case Generation for Autonomous Vehicles with Multimodal Large Language Model
Lu, Q., Ma, M., Wang, Z., Lu, W., Ma, M., Zhao, G., & Feng, S. (2025) · Tsinghua Science & Technology · doi:10.26599/tst.2025.9010178 · Cited by 6
Mei, Y., Nie, T., Sun, J., & Tian, Y. (2025) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc60802.2025.11423790 · Cited by 19
“To explicitly target high-risk cases, some methods formulate scenario generation as a search or optimization problem.”
Scalable Modular Synthetic Data Generation for Advancing Aerial Autonomy
Sabet, M., Palanisamy, P., & Mishra, S. (2022) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2023.104464 · Cited by 6
“To support DR through the manual procedural model to generate realistic and diverse scenes requires significant time and expertise to tune the parameters [14, 18, 19, 22].”
BackgroundMethodReal-to-Synthetic: Generating Simulator Friendly Traffic Scenes from Graph Representation
Tian, Y., Carballo, A., Li, R., & Takeda, K. (2022) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv51971.2022.9827441 · Cited by 6
MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations
Mandlekar, A., Nasiriany, S., Wen, B., Akinola, I., Narang, Y., Fan, L., Zhu, Y., & Fox, D. (2023) · arXiv · doi:10.48550/arxiv.2310.17596 · Cited by 226
Knowledge guided controllable diffusion for enhanced autonomous driving scenarios generation
Shan, C., Guo, L., & Chen, H. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.130616 · Cited by 6
Realistic Urban Traffic Generator Using Decentralized Federated Learning for the SUMO Simulator
Bazán-Guillén, A., Beis-Penedo, C., Cajaraville-Aboy, D., Bautista, P. B., Redondo, R. P. D., Llopis, L. J. D. L. C., Vilas, A. F., Igartua, M. A., & Fernández‐Veiga, M. (2025) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2025.3597019 · Cited by 5
“Similarly, SceneGen [29] employs auto-regressive neural networks to generate realistic traffic scenarios without relying on rigid rule-based systems, offering a scalable approach to simulate behaviorally rich urban environments for autonomous vehicles.”
AGENTS-LLM: Augmentative GENeration of Challenging Traffic Scenarios with an Agentic LLM Framework
Yao, Y., Bhatnagar, S., Mazzola, M., et al. (2025) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros60139.2025.11246348 · Cited by 5
“However, they often lack fine-grained control over the output and generating novel scenarios [7], [8] from scratch can introduce a distributional shift from the original training scenes which undermines the validity of evaluations especially for learning-based planners.”
DragTraffic: Interactive and Controllable Traffic Scene Generation for Autonomous Driving
Wang, S., Sun, G., Ma, F., Hu, T., Qin, Q., Song, Y., Zhu, L., & Liang, J. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10801623 · Cited by 9
“Other researchers have looked at generating challenging scenes in a more flexible way, such as SceneGen [7] and TrafficGen [8], which pro-pose building scenes in two stages: vehicle placement and trajectory generation.”
Self-citationMahjourian, R., Mu, R., Likhosherstov, V., Mougin, P., Huang, X., Messias, J. V., & Whiteson, S. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10611115 · Cited by 10
“Most learning-based scene generation methods mainly generate the initial states for the agents [5], [6].”
BackgroundMethodHuang, P., Ding, W., Stoler, B., et al. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11127319 · Cited by 5
WorldGen: A Large Scale Generative Simulator
Singh, C. D., Kumari, R., Fermuller, C. M., Sanket, N. J., & Aloimonos, Y. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10160861 · Cited by 5
“Since WorldGen has the ability to export these parameters, neural autoregressive models such as [58] can be deployed for realistic traffic distribution.”
Sheng, Z., Huang, Z., Qu, Y., Leng, Y., & Chen, S. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00364 · Cited by 4
“For instance, SceneGen [36] introduces a neural autoregressive model that sequentially generates the initial states of traffic agents to compose a static scene.”
Modelling Competitive Behaviors in Autonomous Driving Under Generative World Model
Qiao, G., Quan, G., Qu, R., & Liu, G. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72761-0_2 · Cited by 4
“To improve realism, learning-based approaches [21,39] predict or simulate traffic behavior by learning from a realistic traffic dataset.”
Vectorized Scenario Description and Motion Prediction for Scenario-Based Testing
Winkelmann, M., Vasconi, C., & Müller, S. (2023) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv55152.2023.10186626 · Cited by 4
Learning from Demonstrations of Critical Driving Behaviours Using Driver's Risk Field
Du, Y., Acerbo, F. S., Kober, J., et al. (2023) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2023.10.1376 · Cited by 4
Continuous Locomotive Crowd Behavior Generation
Bae, I., Lee, J., & Jeon, H. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02088 · Cited by 3
“Recent advances in generative models have enabled the synthesis of realistic traffic scenes [30, 60, 93, 161, 162].”
Robust Fast Adaptation from Adversarially Explicit Task Distribution Generation
Wang, Q., Lv, Y., Mao, Y., Qu, Y. P., Xu, Y., & Ji, X. (2025) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3690624.3709337 · Cited by 4
Deep Learning Approach for Realistic Traffic Video Changes Across Lighting and Weather Conditions
Alam, M. S., Parmar, S. H., Martens, M., & Bazilinskyy, P. (2025) · International Conference on Information and Computer Technologies (ICICT) · doi:10.1109/icict64582.2025.00034 · Cited by 3
“Similarly, Tan et al. [21] attempted to create realistic scenes using GANs.”
Enhancing Autonomous Driving Safety with Collision Scenario Integration
Wang, Z., Lan, S., Sun, X., Chang, N., Li, Z., Yu, Z., & Álvarez, J. M. (2025) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros60139.2025.11246713 · Cited by 10
“SceneGen [39] uses auto-regressive models for scenario generation, while TrafficGen [23] applies modules for agent initialization and motion prediction.”
Safety-Critical Traffic Simulation with Adversarial Transfer of Driving Intentions
Huang, Z., Gao, X., Zheng, G., Wen, L., Yang, X., & Sun, X. L. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11128569 · Cited by 2
“Some pi-oneering research [11]–[13] employs end-to-end deep neural networks, like GNNs [14] and Transformer [15], to learn to capture complex interactions in traffic scenes and generate realistic agent trajectories from large-scale real-world data.”
Testing Scenario Identification for Automated Vehicles Based on Deep Unsupervised Learning
Liu, S., Ren, F., Li, P., et al. (2023) · World Electric Vehicle Journal · doi:10.3390/wevj14080208 · Cited by 2
Meng, Q., Zhang, L., Zhao, C., et al. (2024) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2024.3385792 · Cited by 2
Eroglu, S., Voigt, A., Weyers, B., et al. (2024) · IEEE Conference on Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW) · doi:10.1109/vrw62533.2024.00040 · Cited by 2
“Real public data sources [12,15,26], procedural generation [28,34], learning-based [13,21,33] or user-driven [6,18,38] methods are used to address these needs.”
MethodTowards Interactive Autonomous Vehicle Testing: Vehicle-Under-Test-Centered Traffic Simulation
Liu, Y., Zhao, X., & Sun, J. (2024) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc58415.2024.10920277 · Cited by 4
“Deep learning (DL) models emerge as a viable methodology in addressing these challenges, which can incorporate implicit interactive relationships among traffic participants [4], [16]–[18].”
AutoSceCraft: Generate Various Driving Scenarios from Scratch for Autonomous Driving Systems
Lan, W., Liu, J., Yuan, B., & Yao, X. S. (2025) · Tsinghua Science & Technology · doi:10.26599/tst.2025.9010045 · Cited by 2
Semantically Controllable Scene Generation with Guidance of Explicit Knowledge.Preprint
Ding, W., Li, B., Eun, K. J., et al. (2021) · arXiv (Cornell University) · Cited by 2
CausalAF: Causal Autoregressive Flow for Goal-Directed Safety-Critical Scenes Generation.Preprint
Ding, W., Lin, H., Li, B., et al. (2021) · arXiv (Cornell University) · Cited by 2
Self-driving simulation scene reconstruction using self-supervised depth completion
Zhang, Y., Zheng, X., & Liang, H. (2022) · International Conference on Automation, Control and Robotics Engineering (CACRE) · doi:10.1109/cacre54574.2022.9834190 · Cited by 1
Li, Y., Hou, Z., & Liu, H. (2024) · IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom) · doi:10.1109/trustcom63139.2024.00243 · Cited by 1
“For example, [7] utilizes LSTM models to generate vehicle trajectories.”
SDOA: A Synthetic Dataset for Occlusion-Aware and Amodal Perception in Autonomous Driving
Xia, Y. (2025) · International Conference on Advanced Algorithms and Control Engineering (ICAACE) · doi:10.1109/icaace65325.2025.11019488 · Cited by 1
“Recent research integrates machine learning and rule-based methods to automate scene generation [8]–[10], but challenges remain in controllability and interpretability.”
Chen, Y., Jiang, H., & Sun, T. (2025) · Symmetry · doi:10.3390/sym17091503 · Cited by 1
“Considering that the GAN and VAE struggle with the problem of temporal interaction, Tan [79] used an autoregressive model to characterize the dependence features between historical data, and by recursively learning the conditional probability of historical trajectories, Tan generated hazard…”
SceneDM: Consistent Diffusion Models for Coherent Multi-agent Trajectory Generation
Guo, Z., Gao, X., Zhou, J.-L., et al. (2025) · 2022 IEEE 18th International Conference on Automation Science and Engineering (CASE) · doi:10.1109/case58245.2025.11163744 · Cited by 1
SAGA: Generating Extreme Scenarios for Autonomous Driving via Adversarial Perturbations
Wang, Y., Wu, Q., Shi, X., Li, Y. X., & Li, K. (2025) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2025.3640578 · Cited by 1
“Recent learning-based approaches, including SceneGen [14], Tra ffi cSim [4], Tra ffi cGen [17], and Di ff Scene [18], excel at capturing realistic tra ffi c behaviors from real-world data.”
STGT-Gen: Spatio-Temporal Graph Transformer for Multi-Vehicle Traffic Scenario Generation
Qin, X., Lu, C., Wei, Y., Fan, S., Song, Z., & Gong, J. (2025) · SAE technical papers on CD-ROM/SAE technical paper series · doi:10.4271/2025-01-7316 · Cited by 1
Adversarial and Reactive Traffic Entities for Behavior-Realistic Driving Simulation: A Review
Ransiek, J., Reis, P., Schürmann, T., & Sax, E. (2025) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc58881.2025.11342478 · Cited by 3
“…[92], [99], [100], [101] Scene-centric [6], [13], [67], [69], [70], [71], [73], [78], [79], [80], [81] Scene-centric Scene-centric [14], [102], [103], [104], [105], [106], [107] sampling framework for controlled trajectory generation, while [83] uses a conditional diffusion model for…”
IntersectioNDE: Learning Complex Urban Traffic Dynamics Based on Interaction Decoupling Strategy
Lin, E., Yang, Z., Lu, Q., et al. (2025) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc60802.2025.11423034 · Cited by 1
“From the perspective of generative model architecture, this includes deep Autoregressive Models [22], [23], Variational Autoen-coders (VAE) [24], [25], Generative Adversarial Networks (GAN) [26], [27], and Diffusion Models [28]–[30].”
ARISE – Adaptive Refinement and Iterative Scenario Engineering
Poddubnyy, K., Vozniak, I., Бурмистров, И., et al. (2026) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv66570.2026.11623826 · Cited by 1
“SceneGen [17] learns to sample realistic but abstract traffic scenes from real data, and TrafficGen [18] applies a diffusion model to generate diverse traffic configurations conditioned on map context.”
NeuralFloors: Conditional Street-Level Scene Generation From BEV Semantic Maps via Neural Fields
Muşat, V., Martini, D. D., Gadd, M., et al. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3356793 · Cited by 1
“As a result, research in this area has gained popularity in both academia [1] and industry, where there has been much focus on building digital twins endowed with the diversity and high fidelity of the real world they are emulating.”
BackgroundScene-Extrapolation: Generating Interactive Traffic Scenarios
Zipfl, M., Schütt, B., & Zöllner, J. M. (2024) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv55156.2024.10588703 · Cited by 1
“Tan et al. propose an autoregressive approach, SceneGen [15], to address the limitations of manual and rule-based methods in representing weld complexity.”
MethodTian, H., Wang, X., & Ma, X. (2024) · International Conference on Electronic Information Engineering, Big Data and Computer Technology (EIBDCT) · doi:10.1117/12.3031014 · Cited by 1
“SimNet [6] enhances the generation of diverse simulations by initializing states through a generative model conditioned on semantic maps.”
A Survey of Learning Techniques for Virtual Scene Generation
Ayyildiz, D. V., Alnaser, A. J., Taj, S., Zakaria, M., & Jaimes, L. G. (2024) · SAE International Journal of Connected and Automated Vehicles · doi:10.4271/12-08-02-0017 · Cited by 2
Abouelazm, A., Weinstein, T., Joseph, T., Schörner, P., & Zöllner, J. M. (2025) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv64158.2025.11097458 · Cited by 1
“Restricting nodes to valid road positions [37], [38] ensures that all generated scenarios remain feasible by enabling a proper placement of road-bound NPCs, such as vehicles and motorcycles, as well as roadside obstacles like parked cars or trash containers, without requiring additional masking.”
A Modular Co-Simulation Framework with Open Source Software and Automotive Standards
Salles, D., Lang, L., Kehrer, M., & Reuss, H.-C. (2022) · Proceedings · doi:10.1007/978-3-658-37011-4_18 · Cited by 1
A map-based model-driven testing framework for automated driving systems
Tang, Y. (2022) · doi:10.32657/10356/164257 · Cited by 1
Semantically Controllable Generation of Physical Scenes with Explicit Knowledge.Preprint
Ding, W., Li, B., Eun, K. J., et al. (2021) · arXiv (Cornell University)
Accident Scenario Generation using Driver Behavior Model
Jawad, A., & Whitehead, J. (2024) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc58415.2024.10920274
“Data-driven methods utilize techniques such as direct sampling [34], [6] and density estimation [35], [5], [36], [37], [38], [39] to generate scenarios by re-simulating, perturbing, or sampling from distributions learned from collected datasets.”
Riegl, P., Sekaran, K. C., & Botsch, M. (2024) · IEEE International Conference on Vehicular Electronics and Safety (ICVES) · doi:10.1109/icves61986.2024.10927934
“The works in [13] and [14] use an autoregressive model to generate starting configurations of traffic scenarios.”
Cao, M., Duan, W., Huo, C., et al. (2025) · AIP Advances · doi:10.1063/5.0272210
The Application of Generative Adversarial Network on Data Augmentation for Target Selection
Liu, Y., Wang, Z., Xu, Q., et al. (2025) · International Journal of Automotive Technology · doi:10.1007/s12239-025-00261-5
“Tan et al. (2021) generated realistic traffic scenes without relying on heuristics, accurately modeling real-world distributions.”
HiFiSim:Learning for High-Frequency Traffic Scenario Simulation with Dual-Domain FidelityPreprint
Su, Q., Tang, X., Chen, Z., et al. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5472614
Mammen, M. M., Kayatas, Z., & Bestle, D. (2025) · Future Transportation · doi:10.3390/futuretransp5040159
“To overcome these limitations, recent research turned to deep generative models for data-driven scenario generation, which learn directly from real-world or simulated data [13–21].”
Xie, D., Li, D., Zhang, Y., et al. (2025) · SAE technical papers on CD-ROM/SAE technical paper series · doi:10.4271/2025-99-0133
Benchmarking Generative Models for Traffic Scene Image Synthesis A Comparative Study
Geng, Y., Chen, L., Qiu, B., et al. (2025) · International Conference on Algorithms, Computing and Artificial Intelligence (ACAI) · doi:10.1109/acai68217.2025.11406369
“…images[19]; • Physical implausibility: Unrealistic vehicle interactions (e.g., overlapping cars or floating pedestrians) degrade simulation utility[3]; • Context misunderstanding: Models often fail to interpret specialized terms (e.g., "roundabout right-of-way" or "zebra crossing")[3][12].”
Path Diffuser: Diffusion Model for Data-Driven Traffic Simulator
Lee, D., Karthikeyan, A., Pant, Y. V., & Fischmeister, S. (2025) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc60802.2025.11423013
“SceneControl outperforms popular autoregressive baselines such as those presented in [22].”
Simulation Module of Autonomous Driving
Huang, Y., & Yang, Z. (2026) · Springer eBooks · doi:10.1007/978-981-95-3725-9_10
HABIT: Human Action Benchmark for Interactive Traffic in CARLA
Ramesh, M., Azer, M., & Flohr, F. B. (2026) · IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv61042.2026.00690
“Although they emphasize addressing both appearance and content gaps, particularly in actor behavior generation, the focus has predominantly been on vehicle traffic and sensor realism, not pedestrian-driven dynamics [11].”
Large Language Model-Based Realistic Safety-Critical Driving Video Generation
Fu, Y., Zha, R., Tian, P., & Di, X. (2026) · Transportation Research Record Journal of the Transportation Research Board · doi:10.1177/03611981261445426 · Cited by 2
Zhou, R., Li, S. W., Liu, X., Huang, H., Bian, J., & Wei, Z. (2026) · Transportmetrica A Transport Science · doi:10.1080/23249935.2026.2694539
Metadata
- Authors
- Tan, S., Wong, K., Wang, S., Manivasagam, S., Ren, M., & Urtasun, R.
- Affiliations
- Sun Yat-sen University (CN)
- University of Toronto (CA)
- Venue
- 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2021-06
- Pages
- 892-901
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W3173280621
- Semantic Scholar
- 8b817a28d880e4688956d6e1769475a9f22c7411
- Cited by
- 104
- References
- 64
- Citations by year
- 2026: 10 · 2025: 36 · 2024: 29 · 2023: 17 · 2022: 8 · 2021: 4
- FWCI
- 14.34
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion, cc-by)
- arXiv (Cornell University) (cc-by)
- Topics
- Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Vision and Imaging
- Physical Sciences › Engineering › Control and Systems Engineering › Simulation and Modeling Applications
- Keywords
- autonomous driving simulation, high-definition maps, sensor simulation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
