Enhanced multi agent coordination algorithm for drone swarm patrolling in durian orchards.

Area of Science:

• Robotics and Automation
• Artificial Intelligence
• Agricultural Technology

Background:

• Drone swarms offer potential for efficient agricultural monitoring, but controlling them in complex environments like durian orchards presents challenges.
• Existing algorithms often struggle with dynamic environmental changes and real-time obstacle avoidance, limiting their effectiveness.

Purpose of the Study:

• To propose and validate an enhanced multi-agent swarm control algorithm (EN-MASCA) for efficient drone swarm patrolling in complex durian orchards.
• To improve the flexibility, stability, and adaptability of drone swarms in dynamic and complex environments.

Main Methods:

• Development of a virtual navigator model for real-time path adjustment, obstacle avoidance, and optimization.
• Application of deep reinforcement learning for adaptive path planning and obstacle avoidance.
• Incorporation of biological swarm behavior simulation to optimize flight paths, obstacle avoidance, and swarm stability.

Main Results:

• EN-MASCA demonstrated significant improvements in flight trajectory, stability, swarm consistency, and task completion efficiency compared to traditional methods.
• The virtual navigator model enhanced real-time adaptability to environmental changes.
• Deep reinforcement learning improved the algorithm's learning and optimization capabilities in dynamic settings.

Conclusions:

• EN-MASCA provides an efficient and intelligent solution for collaborative drone patrol operations in durian orchards.
• The algorithm shows significant practical application value and promotion prospects for intelligent agricultural management.