Bridging Intuition and Data: A Unified Bayesian Framework for Optimizing Unmanned Aerial Vehicle Swarm Performance

Area of Science:

• Engineering Management
• Artificial Intelligence
• Robotics

Background:

• The rapid expansion of low-altitude economic ecosystems and Unmanned Aerial Vehicle (UAV) swarm applications necessitates advanced performance evaluation and operational optimization strategies.
• Existing evaluation methods are often inadequate for the dynamic and complex nature of UAV swarms, lacking the ability to integrate diverse performance criteria effectively.

Purpose of the Study:

• To introduce a novel Bayesian Network (BN)-based multicriteria decision-making framework for evaluating and optimizing UAV swarm performance.
• To bridge the gap between subjective expert insights and objective real-time data in UAV swarm management.

Main Methods:

• Development of a Bayesian Network (BN) framework integrating expert intuition and real-time data for multicriteria decision-making.
• Utilization of variance decomposition to establish a bidirectional mapping between expert weights and network probabilistic parameters.
• Validation through comprehensive testing to assess the framework's effectiveness in identifying key performance drivers.

Main Results:

• The proposed BN framework successfully integrates expert knowledge and objective data into a unified model.
• The framework effectively identifies critical performance drivers for UAV swarms, such as environmental awareness, communication, and collaborative decision-making.
• Validation confirms the framework's ability to provide transparent and actionable insights for engineering managers.

Conclusions:

• The developed BN framework offers a transparent and adaptive tool for engineering managers overseeing UAV swarm systems.
• The framework facilitates informed resource allocation, technology adoption, and enhances the overall operational effectiveness of complex UAV swarms.
• This approach provides a robust solution for the performance evaluation and optimization challenges posed by growing UAV swarm applications.