Efficacy assessment for multi-vehicle formations based on data augmentation considering reliability

Abstract

Nowadays, aerial vehicle swarm (AVS) formations have been widely applied to military actions. Meanwhile, assessing their efficacy has also received increasing attention due to the significance for offline tactic planning and online formation switching, which places extra emphasis on the combination of empirical knowledge and experimental data, the transparency of assessment process and the explainability of assessment results. In practical engineering, the efficacy data collected from AVS formations are often incomplete due to limited experiments, and unreliable due to multi-source uncertainties, which poses a great challenge to the development of efficacy assessment system that balances reliability, explainability and accuracy. Oriented to the efficacy assessment of AVS formations under incomplete and unreliable efficacy data, the solution based on belief rule base (BRB) is developed. The reliable heterogeneous data augmentation (RHDA) method is first proposed to improve the availability of efficacy data collected from AVS formation with destination configuration by those collected from AVS formation with source configuration. Then, the feature reliability calculation method based on configuration similarity is designed to avoid the interference of subjective intention. Besides, the configuration importance measurement method based on first-order traceability analysis is put forward to provide decision-makers with references for selecting the most important source configuration when multiple ones exist. The feasibility and effectiveness of developed solution are verified by appropriate simulation experiments.