Realistic Unmanned Aerial Vehicle Simulation: A Comprehensive Approach

Abstract

This paper presents a comprehensive software architecture that facilitates the testing and deployment of autonomous missions in the context of Unmanned Aerial Vehicle (UAV) simulation. Our approach integrates and enhances several tools, including the Robot Operating System (ROS), to create a robust environment for evaluating autonomous UAV operations and gathering data. The core of our simulation framework leverages Unreal Engine and the AirSim plugin to model UAV dynamics and generate sensor data. Various modifications have been implemented to enhance sensor fidelity and enable realistic movement of diverse world objects, achieving a higher degree of realism. Designed for high configurability, the proposed architecture allows for flexible adaptation to a wide range of scenarios and requirements. Thus, this simulation environment supports rigorous testing and development of autonomous UAV systems, presenting a valuable platform for researchers and developers in the field of aerial robotics.