AeroSim: An Open Simulator for Aerospace Engineering and Physical AI Research

We introduce AeroSim, an open-source simulator, for aerospace engineering and physical AI research, that has been built from the ground up as an integrated simulation system for aerospace engineering workflows and a testbed for developing, training, and validating physical AI systems. At its core, AeroSim leverages highly realistic flight simulation, FMU integration, multi-renderer support, real-time communications, and robust input handling to deliver a simulation platform that is not only highly accurate but also scalable for advanced deep learning. By uniting aerospace simulation methods with modern computational paradigms, AeroSim lays the groundwork for training world models, control algorithms, vision-action models, and deep reinforcement learning (RL) agents tailored for aerospace engineering and physical AI integration.