What question did this study set out to answer?

The research aims to develop an advanced multi-agent reinforcement learning algorithm for pursuit-evasion scenarios involving UAVs.

March 25, 2026AerospaceOpen Access

EO-MADDPG: An Improved Reinforcement Learning Approach for Multi-UAV Pursuit–Evasion Games

View Full Paper

Ask AI

Bookmark

Authors

Novel algorithm EO-MADDPG enhances pursuit–evasion outcomes in multi-UAV settings, indicating superior performance.

The research aims to develop an advanced multi-agent reinforcement learning algorithm for pursuit-evasion scenarios involving UAVs.
Introduced EO-MADDPG algorithm combining expert knowledge with real-time data sampling and opponent action prediction.
Implemented multi-UAV formation control and encirclement strategies using consensus algorithms and Apollonius circle guidance.
Optimized network training by integrating opponent actions under a fixed policy for better prediction.
EO-MADDPG achieved success rates of 99.9 ± 0.3% in three vs. one and 97.5 ± 1.4% in three vs. two scenarios.
Outperformed the baseline MADDPG, which had success rates of 72.7 ± 6.0% and 64.4 ± 34.4%.
Ablation studies showed enhanced training stability and improved interpretability.

Wang et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark