Research of UAV 3D path planning based on improved Dwarf mongoose algorithm with multiple strategies.

Area of Science:

• Robotics
• Artificial Intelligence
• Optimization Algorithms

Background:

• Unmanned Aerial Vehicles (UAVs) face challenges in complex operational environments.
• Efficient and rapid path planning is crucial for UAV adaptability and cost reduction.

Purpose of the Study:

• To propose a novel 3D UAV path planning algorithm based on an improved Dwarf Mongoose Optimization (DMO).
• To enhance UAV adaptability, speed, efficiency, and reduce operational costs in complex environments.

Main Methods:

• Implemented a chaos mapping-based opposition-based learning strategy for improved initial population distribution and global search.
• Introduced a golden sine function with nonlinear weights to balance exploration and exploitation and avoid local optima.
• Incorporated a differential mutation strategy to enhance population diversity and escape local optima.
• Validated the improved algorithm (CDMOS) through ablation experiments and Wilcoxon rank-sum tests on benchmark functions.

Main Results:

• The improved CDMOS algorithm demonstrated superior optimization performance, convergence precision, and stability compared to the original DMO.
• Achieved an average improvement of 53.5% in convergence accuracy and 35.1% in solution stability across 29 benchmark functions.
• In 3D path planning simulations, CDMOS reduced path length by 46.0% and smoothness cost by 93.4%, while maintaining low obstacle cost.

Conclusions:

• The CDMOS algorithm effectively enhances UAV robustness, adaptability, and real-time performance in complex 3D path planning scenarios.
• The generated flight paths are optimized for stability and efficiency, making the algorithm suitable for demanding mission requirements.