Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Tactics2D: A Highly Modular and Extensible Simulator for Driving Decision-making

Yueyuan Li, Songan Zhang, Mingyang Jiang, Xingyuan Chen, Yeqiang Qian, Chunxiang Wang, Ming Yang

TL;DR

Tactics2D tackles the need for diverse, interactive driving scenarios by delivering a highly modular, open-source driving decision-making simulator. Its design centers on a pluggable architecture with parsers for maps and trajectories, a map generator, a central scenario manager, sensor models, and configurable behavior and physics models, all supporting log replay and dataset compatibility. Key contributions include built-in environments, flexible import/export of custom trajectories and maps, extensive dataset compatibility, interactive NPCs, multi-agent capabilities, and a lightweight runtime suitable for rapid development and benchmarking. The framework enables researchers to rapidly develop, test, and compare driving decision-making methods with realistic variability, while online materials and future plans (e.g., 3D co-simulation and CARLA/SUMO interfaces) promise broader applicability and community-driven growth.

Abstract

Simulation is a prospective method for generating diverse and realistic traffic scenarios to aid in the development of driving decision-making systems. However, existing simulators often fall short in diverse scenarios or interactive behavior models for traffic participants. This deficiency underscores the need for a flexible, reliable, user-friendly open-source simulator. Addressing this challenge, Tactics2D adopts a modular approach to traffic scenario construction, encompassing road elements, traffic regulations, behavior models, physics simulations for vehicles, and event detection mechanisms. By integrating numerous commonly utilized algorithms and configurations, Tactics2D empowers users to construct their driving scenarios effortlessly, just like assembling building blocks. Users can effectively evaluate the performance of driving decision-making models across various scenarios by leveraging both public datasets and user-collected real-world data. For access to the source code and community support, please visit the official GitHub page for Tactics2D at https://github.com/WoodOxen/Tactics2D.

Tactics2D: A Highly Modular and Extensible Simulator for Driving Decision-making

TL;DR

Tactics2D tackles the need for diverse, interactive driving scenarios by delivering a highly modular, open-source driving decision-making simulator. Its design centers on a pluggable architecture with parsers for maps and trajectories, a map generator, a central scenario manager, sensor models, and configurable behavior and physics models, all supporting log replay and dataset compatibility. Key contributions include built-in environments, flexible import/export of custom trajectories and maps, extensive dataset compatibility, interactive NPCs, multi-agent capabilities, and a lightweight runtime suitable for rapid development and benchmarking. The framework enables researchers to rapidly develop, test, and compare driving decision-making methods with realistic variability, while online materials and future plans (e.g., 3D co-simulation and CARLA/SUMO interfaces) promise broader applicability and community-driven growth.

Abstract

Simulation is a prospective method for generating diverse and realistic traffic scenarios to aid in the development of driving decision-making systems. However, existing simulators often fall short in diverse scenarios or interactive behavior models for traffic participants. This deficiency underscores the need for a flexible, reliable, user-friendly open-source simulator. Addressing this challenge, Tactics2D adopts a modular approach to traffic scenario construction, encompassing road elements, traffic regulations, behavior models, physics simulations for vehicles, and event detection mechanisms. By integrating numerous commonly utilized algorithms and configurations, Tactics2D empowers users to construct their driving scenarios effortlessly, just like assembling building blocks. Users can effectively evaluate the performance of driving decision-making models across various scenarios by leveraging both public datasets and user-collected real-world data. For access to the source code and community support, please visit the official GitHub page for Tactics2D at https://github.com/WoodOxen/Tactics2D.
Paper Structure (29 sections, 1 figure, 2 tables)

This paper contains 29 sections, 1 figure, 2 tables.

Table of Contents

  1. Introduction
  2. Design of Tactics2D
  3. Parser
  4. Map parser
  5. Trajectory parser
  6. Map generator
  7. Scenario manager
  8. Sensor model
  9. Map paradigm
  10. Traffic participants in scenario manager
  11. Behavior model
  12. Physics model
  13. Updaters in scenario manager
  14. Render updater
  15. Physics updater
  16. ...and 14 more sections

Figures (1)

  • Figure 1: The structure of Tactics2D. The blue blocks are the executable code modules that the users can customize. The orange blocks are the data conveyed through the simulator.