AirSim is a simulator for autonomous vehicles built on Unreal Engine and Unity that enables physically and visually realistic testing of drones, cars, and other platforms.
The tool addresses the challenge of developing and testing autonomous vehicle algorithms without access to real hardware or safe testing environments. It provides software-in-the-loop simulation with flight controllers like PX4 and ArduPilot, as well as hardware-in-the-loop support for PX4. The simulator runs as an Unreal plugin that can be integrated into any Unreal environment, with an experimental Unity plugin also available. AirSim exposes platform-independent APIs through RPC that allow programmatic control and data retrieval in multiple languages including C++, Python, C#, and Java, enabling researchers to write and test code in simulation before deploying to real vehicles.
Developers should choose AirSim if they are conducting AI research involving deep learning, computer vision, or reinforcement learning for autonomous systems. It suits projects requiring realistic visual and physical simulation across Windows, Linux, and macOS. The tool is particularly valuable for teams practicing transfer learning, as the same API-based code can run in simulation and on actual hardware. Manual control via remote control for drones and keyboard for cars is also supported for interactive testing.
The project maintains a substantial user base, with almost all open issues raised by outside adopters rather than the core team. Responses to issues and pull requests typically arrive within one to two weeks. Work in the issue tracker centers on questions from users, bug reports, and image-related functionality.
