系统工程与电子技术

• 系统工程 • 上一篇    下一篇

基于多目标优化算法的多无人机协同航迹规划

周德云, 王鹏飞, 李枭扬, 张堃   

  1. 西北工业大学电子信息学院, 陕西 西安 710129
  • 出版日期:2017-03-23 发布日期:2010-01-03

Cooperative path planning of multi-UAV based on multi-objective optimization algorithm

ZHOU Deyun, WANG Pengfei, LI Xiaoyang, ZHANG Kun   

  1. College of Electronics and Information, Northwestern Polytechnical University, Xi’an 710129, China
  • Online:2017-03-23 Published:2010-01-03

摘要:

多无人机协同航迹规划是无人机协同作战的关键技术之一。本文提出的一种针对多无人机协同航迹规划的多目标优化算法,即协同非支配排序进化算法(cooperated nondominated sorting genetic algorithms II,CO-NSGA II),针对多架无人机的航迹距离、安全性、时间以及空间的协同性进行规划。运用多目标优化算法,克服了传统航迹规划中需要为各目标函数取权值的不足,并且可以生成多组可供选择的解。同时引入协同进化策略,将各无人机的航迹规划视作子种群,各子种群间进行合作,子种群内采用非支配排序进化算法(non-dominated sorting genetic algorithms II,NSGA II)进行独立优化。考虑到各机间的协同约束,用时间空间协同系数替代传统算法中的“拥挤距离”参数。仿真结果表明通过本文算法能够有效实现多无人机协同航迹规划。

Abstract:

Cooperative path planning of multiple unmanned aerial vehicle(multi-UAV)is one of the key technologies of UAV cooperative engagement. A multi-objective optimization algorithm for cooperative path planning of multi-UAV, which is named as cooperated nondominated sorting genetic algorithms II(CO-NSGAII) is proposed for planning track distance, safety, spatial and time cooperativity of multi-UAV. By using the multi-objective optimization algorithm, the deficiency of taking weight for every objective function in the traditional path planning is overcame, and can get multiple alternative results. Meanwhile, by introducing the co-evolution strategy, the path planning of each UAV is treated as sub population, the best individual is used to cooperate between sub populations, and multiple objectives are optimized by non-dominated sorting genetic algorithms II(NSGA II) respectively in each sub population. Considering spatial and time constraints of UAV, the parameter of “crowding distance” in traditional algorithm is replaced by the parameter of spatial and time cooperativity. The simulation results show that the proposed algorithm can achieve cooperative path planning of multi-UAV effectively.