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

Simulators generate diverse and realistic traffic scenarios to boost the development of autonomous driving systems. However, existing simulators often fall short in scenario diversity and interactive behavior models for traffic participants. This deficiency underscores the need for a flexible, reliable, user-friendly open-source simulator. Addressing this challenge, Tactics2D provides a highly modular and extensive framework for traffic scenario construction, encompassing road elements, traffic regulations, behavior models, physics simulations for vehicles, and event detection mechanisms. By integrating numerous popular algorithms and models, Tactics2D empowers users to customize driving scenarios and evaluate model performance across various scenarios by leveraging both public datasets and user-collected real-world data. This letter results from discussions at several IEEE T-IV's Decentralized and Hybrid Workshops on Scenarios Engineering for Smart Mobility.