Awesome-Traffic-Agent-Scene-Simulation-For-Autonomous-Driving

Papers related to data-driven traffic agent or traffic scene simulation for autonomous driving, including:

• learning based traffic agent model(single agent and multi-agents)
• traffic scene generation
• advesarial trajectory and traffic scene generation

Some papers focus on a more general traffic agent simulator, while some papers focus on safety-critical behavior or scenario in particular. Welcome to contribute :)

TODO

• add genre of project
• add tag of methods(eg, RL/IL) and publication source

Workshops

• IV 2024 Workshop SAFE-DRIVE: Data-Driven Simulations and Multi-Agent Interactions for Autonomous Vehicle Safety Website

• CVPR2024 Workshop on Data-Driven Autonomous Driving Simulation. Website

• IROS2023 Workshop on Traffic Agent Modeling for Autonomous Driving Simulation. Website

Papers

Listed by order of time(not strictly)

• GPD-1: Generative Pre-training for Driving. arXiv Project

• Closed-Loop Supervised Fine-Tuning of Tokenized Traffic Models. arXiv Project Code

• DiffRoad: Realistic and Diverse Road Scenario Generation for Autonomous Vehicle Testing. arXiv

• FREA: Feasibility-Guided Generation of Safety-Critical Scenarios with Reasonable Adversariality. arXiv Project Code

• AdvDiffuser: Generating Adversarial Safety-Critical Driving Scenarios via Guided Diffusion. arXiv

• Improving Agent Behaviors with RL Fine-tuning for Autonomous Driving. Paper

• Traffic Scene Generation from Natural Language Description for Autonomous Vehicles with Large Language Model. arXiv Project Code

• Language-Driven Interactive Traffic Trajectory Generation. arXiv Github

• Data-driven Diffusion Models for Enhancing Safety in Autonomous Vehicle Traffic Simulations. arXiv

• Learning to Drive via Asymmetric Self-Play. arXiv

• SEAL: Towards Safe Autonomous Driving via Skill-Enabled Adversary Learning for Closed-Loop Scenario Generation. arXiv Project Github

• Controllable Traffic Simulation through LLM-Guided Hierarchical Chain-of-Thought Reasoning. arXiv

• ReGentS: Real-World Safety-Critical Driving Scenario Generation Made Stable. arXiv

• Realistic Extreme Behavior Generation for Improved AV Testing. arXiv

• Adversarial and Reactive Traffic Agents for Realistic Driving Simulation. arXiv

• Promptable Closed-loop Traffic Simulation. arXiv Project Github

• TrafficGamer: Reliable and Flexible Traffic Simulation for Safety-Critical Scenarios with Game-Theoretic Oracles arXiv Project Github

• GPUDrive: Data-driven, multi-agent driving simulation at 1 million FPS. arXiv Github

• DriveArena: A Closed-loop Generative Simulation Platform for Autonomous Driving. arXiv Project Github

• Tactics2D: A Reinforcement Learning Environment Library with Generative Scenarios for Driving Decision-making. arXiv Github

• KiGRAS: Kinematic-Driven Generative Model for Realistic Agent Simulation. arXiv Project

• GUMP: Solving Motion Planning Tasks with a Scalable Generative Model. arXiv Github

• LCSim: A Large-Scale Controllable Traffic Simulator. Project arXiv Github

• Model Predictive Simulation Using Structured Graphical Models and Transformers. arXiv

• NAVSIM: Data-Driven Non-Reactive Autonomous Vehicle Simulation and Benchmarking. arXiv

• RACL: Risk Aware Closed-Loop Agent Simulation with High Fidelity. Paper

• SAFE-SIM: Safety-Critical Closed-Loop Traffic Simulation with Controllable Adversaries arXiv Github

• KnowMoformer: Knowledge-Conditioned Motion Transformer for Controllable Traffic Scenario Simulation. Paper

• GOOSE: Goal-Conditioned Reinforcement Learning for Safety-Critical Scenario Generation. arXiv

• Text-to-Drive: Diverse Driving Behavior Synthesis via Large Language Models. Project arXiv

• BehaviorGPT: Smart Agent Simulation for Autonomous Driving with Next-Patch Prediction. arXiv

• ChatScene: Knowledge-Enabled Safety-Critical Scenario Generation for Autonomous Vehicles. arXiv

• SMART: Scalable Multi-agent Real-time Simulation via Next-token Prediction. Project arXiv Github

• SceneControl: Diffusion for Controllable Traffic Scene Generation. Project PDF Poster

• Towards Interactive Autonomous Vehicle Testing: Vehicle-Under-Test-Centered Traffic Simulation. arXiv

• TorchDriveEnv: A Reinforcement Learning Benchmark for Autonomous Driving with Reactive, Realistic, and Diverse Non-Playable Characters. arXiv Github

• UniGen: Unified Modeling of Initial Agent States and Trajectories for Generating Autonomous Driving Scenarios. arXiv

• TSDiT: Traffic Scene Diffusion Models With Transformers. arXiv

• MRIC: Model-Based Reinforcement-Imitation Learning with Mixture-of-Codebooks for Autonomous Driving Simulation. arXiv

• Scene-Extrapolation: Generating Interactive Traffic Scenarios. arXiv

• Dragtraffic: A Non-Expert Interactive and Point-Based Controllable Traffic Scene Generation Framework. Project arXiv Github

• Enhancing Autonomous Vehicle Training with Language Model Integration and Critical Scenario Generation. arXiv

• Versatile Scene-Consistent Traffic Scenario Generation as Optimization with Diffusion. Project arXiv

• WcDT: World-centric Diffusion Transformer for Traffic Scene Generation. arXiv Github

• CtRL-Sim: Reactive and Controllable Driving Agents with Offline Reinforcement Learning. Project arXiv Github

• SLEDGE: Synthesizing Simulation Environments for Driving Agents with Generative Models. arXiv Github

• CaDRE: Controllable and Diverse Generation of Safety-Critical Driving Scenarios using Real-World Trajectories. arXiv

• Dream to Drive with Predictive Individual World Model. Project Github

• HMSim: A Hierarchical Multi-Agent Learning-Based Simulator For Urban Driving Scenarios. Paper Project

• LitSim: Conflict-aware Policy for Long-term Interactive Traffic Simulation. arXiv

• LimSim++: A Closed-Loop Platform for Deploying Multimodal LLMs in Autonomous Driving. arXiv Project Github

• Editable Scene Simulation for Autonomous Driving via Collaborative LLM-Agents. arXiv Project Github

• OASim: an Open and Adaptive Simulator based on Neural Rendering for Autonomous Driving. arXiv Project Github

• Trajeglish: Learning the Language of Driving Scenarios. Website arXiv

• Controllable Safety-Critical Closed-loop Traffic Simulation via Guided Diffusion. arXiv

• RealGen: Retrieval Augmented Generation for Controllable Traffic Scenarios. Project arXiv

• RITA: Boost Driving Simulators with Realistic Interactive Traffic Flow. arXiv

• SceneDM: Scene-level Multi-agent Trajectory Generation with Consistent Diffusion Models. arXiv Project

• Data-driven Traffic Simulation: A Comprehensive Review. arXiv

• Scenario Diffusion: Controllable Driving Scenario Generation With Diffusion. arXiv

• Learning Realistic Traffic Agents in Closed-loop. arXiv

• Language-Guided Traffic Simulation via Scene-Level Diffusion. arXiv Github

• Guided Conditional Diffusion for Controllable Traffic Simulation Guided Conditional Diffusion for Controllable Traffic Simulation. arXiv Github

• Waymax: An Accelerated, Data-Driven Simulator for Large-Scale Autonomous Driving Research. arXiv Github

• TrafficBots: Towards World Models for Autonomous Driving Simulation and Motion Prediction. arXiv Github

• DriveSceneGen: Generating Diverse and Realistic Driving Scenarios from Scratch. arXiv Project Github

• TrafficMCTS: A Closed-Loop Traffic Flow Generation Framework with Group-Based Monte Carlo Tree Search. arXiv

• SurrealDriver: Designing Generative Driver Agent Simulation Framework in Urban Contexts based on Large Language Model. arXiv

• Reinforcement Learning with Human Feedback for Realistic Traffic Simulation. arXiv

• From Model-Based to Data-Driven Simulation: Challenges and Trends in Autonomous Driving. arXiv

• Language Conditioned Traffic Generation. arXiv Project Github

• A Survey on Safety-Critical Driving Scenario Generation -- A Methodological Perspective. arXiv

• MIXSIM: A Hierarchical Framework for Mixed Reality Traffic Simulation. Paper

• Generating Driving Scenes with Diffusion. arXiv

• TransWorldNG: Traffic Simulation via Foundation Model. arXiv Github

• AdvSim: Generating Safety-Critical Scenarios for Self-Driving Vehicles. arXiv

• Traffic-Aware Autonomous Driving with Differentiable Traffic Simulation. arXiv

• Editing Driver Character: Socially-Controllable Behavior Generation for Interactive Traffic Simulation. arXiv

• TrafficGen: Learning to Generate Diverse and Realistic Traffic Scenarios. arXiv Project Github

• Guided Conditional Diffusion for Controllable Traffic Simulation. arXiv

• InterSim: Interactive Traffic Simulation via Explicit Relation Modeling. arXiv Project Github

• BITS: Bi-level Imitation for Traffic Simulation. arXiv Blog Github

• Symphony: Learning Realistic and Diverse Agents for Autonomous Driving Simulation. arXiv

• KING: Generating Safety-Critical Driving Scenarios for Robust Imitation via Kinematics Gradients. arXiv Project Github

• TrajGen: Generating Realistic and Diverse Trajectories with Reactive and Feasible Agent Behaviors for Autonomous Driving. arXiv Github

• Generating Useful Accident-Prone Driving Scenarios via a Learned Traffic Prior. arXiv Project Github

• Learning to Collide: An Adaptive Safety-Critical Scenarios Generating Method. arXiv

• SceneGen: Learning to Generate Realistic Traffic Scenes. arXiv

• TrafficSim: Learning to Simulate Realistic Multi-Agent Behaviors. arXiv

• Behaviorally Diverse Traffic Simulation via Reinforcement Learning. arXiv

Acknowledgement

• End-to-end Autonomous Driving
• Awesome-Trajectory-Motion-Prediction-Papers