Tencent has recently launched TAD Sim, an autonomous driving simulation software designed specifically for the development and validation of intelligent driving systems. This cross-platform distributed simulation system is built on Tencent's professional gaming engine and centimeter-level digital twin 3D reconstruction technology. It integrates industrial-grade vehicle dynamics models, physics-level sensor simulations, and data-driven AI traffic flow models, enabling closed-loop simulation validation of all modules on actual vehicles.

WeChat Screenshot_20240813102907.png

The architecture of TAD Sim includes modules such as data access and scenario construction, simulation core engine, test analysis and evaluation, test task management, and external access interfaces, supporting an extensible and compatible simulation system architecture. This platform service covers all stages of the vehicle validation V-model process, including model-in-the-loop, software-in-the-loop, hardware-in-the-loop, vehicle-in-the-loop, and driver-in-the-loop, supporting scenario and map management, allowing users to customize scenarios and maps through scenario and map editors.

Key features of TAD Sim include high-fidelity scenario reproduction, high-precision vehicle dynamics models, a rich array of high-precision sensor models, data-driven AI traffic flow models, and efficient testing tools, facilitating the achievement of simulation tests equivalent to millions of kilometers driven per day. Additionally, the platform supports the OpenX international standard format, ensuring compatibility of scenario map data and consistency with industry standards.

TAD Sim is designed for application scenarios in autonomous driving simulation testing and evaluation, supporting closed-loop simulation validation of all modules on actual vehicles, including perception, decision-making, and control algorithms. It can be fully integrated with the entire process of autonomous driving cloud toolchain services such as CI/CD, data management, data collection and annotation, and algorithm training. As the product continues to evolve, TAD Sim is expected to expand into broader application scenarios such as intelligent connected vehicle simulation testing, synthetic data generation, and virtual test drives.

Github Address: https://github.com/Tencent/TAD_Sim