Multi-Modal Multi-Purpose Mobility Simulator using Unreal Engine

Abstract

This thesis proposes a new multi-modal simulator for testing mobility S/W, built on Unreal Engine and engaging various modalities including simulation platforms, operational domains, and environmental conditions. Existing testing frameworks often necessitate manual configurations and provide limited modalities in each simulation, lacking interoperability across different simulators and plugins. To address these issues, first, a systemic plugin integration is carried out to allow users to flexibly configure diverse test suites selectively by combining various modules. Second, we introduce an automated evaluation process that systematically varies the parameters of the selected simulation modules to discover critical test scenarios and edge cases, and evaluates the target S/W with respect to appropriate metrics tailored to the target S/W. This automation significantly enhances the simulation’s efficiency in identifying the critical test scenarios for the target S/W. The proposed framework excels in executing high-dimensional and complicated simulation testing, which will greatly facilitate mobility S/W verification and validation.