Design and Development of a Real-Time Software Framework for Emulating Aircraft Sensor Systems for Digital Twins in Unreal Engine 5

Abstract

The goal of this thesis is the creation of a novel Unreal Engine 5 aircraft sensor emulation framework (SenSim) for Digital Twins in Unmanned Aerial Vehicle system validation. The framework's requirements include real-time capability, high sensor data fidelity, and a modular, extendible architecture. Within the thesis, these requirements are expanded upon, and the final implementation is explained. To prove the validity of the framework regarding the stated use case and requirements, a series of tests was conducted. These tests included a Digital Twin flight test, which proved a low closed-loop latency, as well as an offline Software In The Loop test for virtual cameras, showing a substantial improvement in latency when compared to Microsoft's AirSim. Finally, some synthetic sensor data is highlighted, and a validation use case is presented.