What question did this study set out to answer?

To develop and evaluate a scalable multi-agent reinforcement learning algorithm that resolves delayed response and poor coordination in drone swarm dynamic obstacle avoidance.

August 16, 2026Advanced ElectromagneticsOpen Access

Modeling Dynamic Obstacle Avoidance Strategy of Drone Swarms Combined with Multi-Agent Reinforcement Learning

View Full Paper

Ask AI

Bookmark

Authors

XFX. H. FangLishui Vocational and Technical College KCK. ChenRongsheng Petrochemical (China)CRC. H. RenVi Technology (United States)

Discussion

Member takes

Implication

Computational evaluation demonstrates enhanced dynamic obstacle avoidance in drone swarms via decoupled reinforcement learning, indicating robust real-time coordination.

Key Points

To develop and evaluate a scalable multi-agent reinforcement learning algorithm that resolves delayed response and poor coordination in drone swarm dynamic obstacle avoidance.
Designed the Locally-decoupled and Embedding-enhanced Multi-Agent Deep Deterministic Policy Gradient (LDE-MADDPG) algorithm.
Integrated a Graph Attention Network (GAT) to embed variable-length observations, alongside a dual-path critic and a hierarchical multi-objective reward structure.
Evaluated performance across eight static and dynamic obstacle simulation environments and 100 randomized scenarios.
Achieved collision rates of 2.1%–4.2% in static obstacle scenarios and 3.8%–7.2% in dynamic obstacle scenarios, significantly outperforming baseline algorithms.
Attained a 97.5% mission completion rate across 100 randomized dynamic test scenarios.

Cite This Study

Fang et al. (2026) studied this question.

synapsesocial.com/papers/6a8179dcf2fb91fc834ad3e5 https://doi.org/10.7716/aem.v15i3.3514

View Full Paper

Ask AI

Bookmark

Implication

Computational evaluation demonstrates enhanced dynamic obstacle avoidance in drone swarms via decoupled reinforcement learning, indicating robust real-time coordination.

Modeling Dynamic Obstacle Avoidance Strategy of Drone Swarms Combined with Multi-Agent Reinforcement Learning

Discussion

Member takes

Implication

Key Points

Cite This Study

Modeling Dynamic Obstacle Avoidance Strategy of Drone Swarms Combined with Multi-Agent Reinforcement Learning

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider