ERRT-GA: Expert Genetic Algorithm with Rapidly Exploring Random Tree Initialization for Multi-UAV Path Planning

In unmanned aerial vehicle (UAV) path planning, evolutionary algorithms are commonly used due to their ability to handle high-dimensional spaces and wide generality. However, traditional evolutionary algorithms have difficulty with population initialization and may fall into local optima. This paper...

Full description

Saved in:
Bibliographic Details
Published inDrones (Basel) Vol. 8; no. 8; p. 367
Main Authors Xu, Hong, Niu, Zijing, Jiang, Bo, Zhang, Yuhang, Chen, Siji, Li, Zhiqiang, Gao, Mingke, Zhu, Miankuan
Format Journal Article
LanguageEnglish
Published Basel MDPI AG 01.08.2024
Subjects
Algorithms
Collaboration
Computing time
Convergence
Drone aircraft
Efficiency
Energy consumption
Evolutionary algorithms
expert genetic
Genetic algorithms
Heuristic
multi-UAV
Optimization
Path planning
Planning
RRT initialization
Search algorithms
task assignment
Unmanned aerial vehicles
Online AccessGet full text

Cover

More Information
Summary:In unmanned aerial vehicle (UAV) path planning, evolutionary algorithms are commonly used due to their ability to handle high-dimensional spaces and wide generality. However, traditional evolutionary algorithms have difficulty with population initialization and may fall into local optima. This paper proposes an improved genetic algorithm (GA) based on expert strategies, including a novel rapidly exploring random tree (RRT) initialization algorithm and a cross-variation process based on expert guidance and the wolf pack search algorithm. Experimental results on baseline functions in different scenarios show that the proposed RRT initialization algorithm improves convergence speed and computing time for most evolutionary algorithms. The expert guidance strategy helps algorithms jump out of local optima and achieve suboptimal solutions that should have converged. The ERRT-GA is tested for task assignment, path planning, and multi-UAV conflict detection, and it shows faster convergence, better scalability to high-dimensional spaces, and a significant reduction in task computing time compared to other evolutionary algorithms. The proposed algorithm outperforms most other methods and shows great potential for UAV path planning problems.
ISSN:2504-446X
2504-446X
DOI:10.3390/drones8080367