A new method for unmanned aerial vehicle path planning in complex environments

Area of Science:

• Robotics
• Artificial Intelligence
• Aerospace Engineering

Background:

• Traditional Unmanned Aerial Vehicle (UAV) path planning faces challenges including low search efficiency, path unevenness, and limited adaptability in unknown environments.
• Existing methods often struggle to balance global path optimality with real-time local obstacle avoidance.

Purpose of the Study:

• To propose a robust UAV path planning method that addresses the limitations of existing approaches.
• To enhance search efficiency, path smoothness, and adaptability to dynamic, unknown environments.
• To effectively fuse global and local path planning strategies for complex tasks.

Main Methods:

• A double-layer optimization approach combining an enhanced A* algorithm with a modified Dynamic Window Approach (DWA).
• Optimized A* using a neighboring node clip-off rule and an obstacle coverage model for improved node expansion and heuristic function.
• Modified DWA with a new tracking index and dynamic adaptive evaluation function weights for better global-local path fitting and obstacle avoidance.

Main Results:

• The proposed method significantly improves path planning efficiency and smoothness for mobile robots.
• Enhanced real-time obstacle avoidance capabilities and adaptability in unknown environments were demonstrated.
• Successful fusion of A* and DWA algorithms achieved superior performance in complex planning scenarios.

Conclusions:

• The developed double-layer optimization method effectively overcomes the limitations of conventional UAV path planning techniques.
• The approach offers a significant advancement in creating efficient, smooth, and adaptive paths for UAVs.
• This method provides a robust solution for complex path planning tasks, particularly in dynamic and unknown environments.