Unmanned Vehicles: A Survey of Modern Simulators

Abstract. This survey is devoted to popular simulators supporting rough terrain for unmanned vehicles, namely, Gazebo, CARLA, AirSim, NVIDIA Isaac Sim, and Webots. Their main capabilities related to terrain modeling, motion physics, and support for sensors and weather conditions are described. Particular attention is paid to the creation of realistic rough terrain scenes, the complexity of importing real maps, and interaction with other software platforms, such as Robot Operating System (ROS) and artificial intelligence (AI) systems. The main drawbacks of each simulator are analyzed: the labor intensity of creating detailed terrain and vehicle models, the high complexity of integrating real maps, and the dependence on powerful hardware. The survey also notes the complexity of interaction with various software solutions and the required knowledge of 3D modeling. Gazebo and Webots are remarkable for their good integration with ROS but require more effort to work with rough terrain. CARLA and AirSim provide high-quality visualization but have higher requirements for creating landscapes. NVIDIA Isaac Sim stands out for AI simulation support but is resource-intensive. The authors’ experience in mapping vehicle trajectories and orienting in some simulators is presented.

Keywords: unmanned vehicles, simulators, rough terrain, tracked platforms.

Acknowledgments. The research of N.A. Korgin was supported in part by the Russian Science Foundation, project no. 23-29-00681, https://rscf.ru/project/23-29-00681.

PDF (English)

Cite this paper

Makarov, M.I., Korgin, N.A., and Pyzh’yanov, A.A., Unmanned Ground Vehicles: a Survey of Modern Simulators. Control Sciences 1, 2–12 (2025).