Preference-Based Multi-Agent Collision Avoidance for Unmanned Aerial Vehicle Swarms with Multi-Objective Optimization
Yulong Cao, Guhao Zhao, Hao Wang, 赖光明 · Engineered Science · 2026
A preference-based multi-agent collision avoidance algorithm is proposed for unmanned aerial vehicle swarms, achieving 50% collision rate reduction with real-time performance through the integration of strategy selection network, multi-objective evaluator, and bayesian personalized ranking-based preference learning.