SOTM: A Simulation-Oriented Trustworthiness Measurement Method for Intelligent Unmanned System

Mengcen Jiang, Youdi Gong, Guang Yang, Lingzhong Meng, Yuxi Ma

Abstract: The continuous iterative advancements and growing complexity of artificial intelligence have profoundly influenced the operational efficacy of intelligent unmanned systems. Comprehensive evaluation and testing of system trustworthiness can effectively ensure reliability and safety during task execution. This study concentrates on intelligent unmanned systems in representative operational scenarios and presents a simulation-driven trustworthiness measurement methodology (SOTM). First, we systematically define the fundamental trustworthiness concepts through three dimensions: essential definition, characteristic spectrum, and conceptual connotation. Subsequently, we establish a threelevel hierarchical framework (“Trustworthiness-Dimensional Attributes-Metrics”) with dual core components - capability trustworthiness and operational trustworthiness - forming an integrated quantitative assessment system. Furthermore, we implement context-aware measurement customization through task-specific parameter adaptation, enabling workflow-aligned trustworthiness evaluation. Leveraging virtual scenario modeling technology, the proposed methodology is integrated into simulation testing procedures, realizing comprehensive trustworthiness verification in simulated environments. Empirical validation on a simulation platform confirms that our measurement framework and testing approach can effectively assess and validate the trustworthiness of intelligent unmanned systems in characteristic operational tasks.

External IDs:dblp:conf/isssr/JiangGYMM25