Three-Dimensional Path Planning for UAV Based on Multi-Strategy Dream Optimization Algorithm.

Area of Science:

• Artificial Intelligence
• Robotics
• Optimization Algorithms

Background:

• Intelligent optimization algorithms face challenges in UAV 3D path planning, including limited search capability, slow convergence, and local optima.
• Enhancing global search efficiency and accuracy is crucial for effective UAV navigation in complex 3D environments.

Purpose of the Study:

• To propose the multi-strategy optimized dream optimization algorithm (MSDOA) for superior UAV 3D path planning.
• To improve global search efficiency, convergence speed, and accuracy in UAV path planning algorithms.

Main Methods:

• Utilized Bernoulli chaotic mapping for population initialization to increase diversity and search range.
• Incorporated an adaptive perturbation mechanism and lens imaging reverse learning for enhanced exploration and convergence.
• Developed an Adaptive Individual-level Mixed Strategy (AIMS) for flexible searching and improved global capability.

Main Results:

• MSDOA demonstrated superior optimization accuracy, faster convergence, and enhanced robustness compared to other swarm intelligence algorithms.
• Achieved first rank on 28 out of 29 CEC2017 benchmark functions, showcasing strong global search capability.
• In UAV path planning, MSDOA reduced best cost function fitness by 9% and average cost by 12% in complex scenarios.

Conclusions:

• MSDOA significantly outperforms existing algorithms in both benchmark testing and UAV 3D path planning applications.
• The algorithm generates more efficient, smoother, and higher-quality flight paths for unmanned aerial vehicles.
• MSDOA offers a robust and adaptable solution for complex 3D navigation challenges in robotics.