Cosys-Lab/Cosys-AirSim

What it solves

Cosys-AirSim is a high-fidelity simulation framework for autonomous vehicles, including drones and cars, designed to provide physically and visually realistic environments for research and industrial applications. It bridges the gap between simulation and reality by supporting hardware-in-the-loop simulation with popular flight controllers like PX4.

How it works

Built on Unreal Engine, the project is an Unreal plugin that can be integrated into any Unreal environment. It provides extensive API support and simulates various sensors including cameras with distortion features, GPU-accelerated LiDAR, and Echo sensors (sonar/radar). It supports multiple vehicle models, including skid-steer vehicles and a small BoxCar for indoor use, and integrates with ROS, ROS2, and Matlab.

Who it’s for

Researchers and developers working on autonomous systems, robotics, and vehicle control, specifically those needing high-fidelity sensor simulation and ground-truth label generation for training AI models.

Highlights

• Advanced Sensor Simulation: Includes GPU-accelerated LiDAR for higher point density and realistic intensity, and Echo sensors for sonar and radar.
• Ground Truth Generation: Features multi-layer annotation for RGB, greyscale, and texture options, as well as instance segmentation.
• Hardware Integration: Supports hardware-in-the-loop with PX4 flight controllers.
• Extensible API: Provides comprehensive API support with dedicated implementations for Python ROS, C++ ROS2, and a Matlab toolbox.
• Unreal Engine 5 Support: Updated and optimized for Unreal Engine 5.x versions.