iamaisim/ProjectAirSim

What it solves

Project AirSim provides a flexible simulation platform for developing and testing autonomous systems (such as drones, fixed-wing aircraft, and robots) without needing to maintain separate software integrations for different levels of simulation fidelity. It allows developers to switch between a lightweight, headless environment for fast iteration and a high-fidelity 3D environment for visual and sensor testing.

How it works

The platform is organized into three layers:

1. Simulation Libraries: The core infrastructure that defines robot structures, physics, controllers, and sensors.
2. Simulation Host: The environment where the simulation runs. Users can choose the Project AirSim Runtime (a lightweight, engine-independent host for fast physics and API testing) or the Unreal Engine plugin (which provides 3D worlds, high-fidelity rendering, and complex sensors like LiDAR and radar).
3. Client Libraries: Network APIs (available in Python, C++, and via a ROS 2 bridge) that allow autonomy stacks to control vehicles and receive sensor data.

Who it’s for

It is designed for engineers and researchers building autonomy stacks for drones, aircraft, and robots who need a scalable way to test their code across different simulation environments, from CI/CD pipelines to photorealistic 3D scenarios.

Highlights

• Engine Independence: Ability to run without Unreal Engine for faster controller and physics development.
• Multi-Vehicle Support: Capability to simulate swarms of vehicles in shared environments.
• Diverse Physics Models: Supports C++ Fast Physics, JSBSim for fixed-wing aircraft, and Simulink physics models.
• Broad Sensor Suite: Includes simulated cameras, LiDAR, radar, IMU, GPS, barometer, and magnetometer.
• Industry Integrations: Built-in support for PX4 (SITL/HITL), ArduPilot SITL, and ROS 2 Humble.