基于改进梯度策略的多虚拟结构算法的无人机协同控制  

Unmanned aerial vehicles collaborative control method based on improved gradient strategy for multiple virtual structures control

作　　者：陈子强 杨艺[1] 姚雪莲[1] CHEN Ziqiang;YANG Yi;YAO Xuelian(School of Automative and Traffic Engineering,jiangsu University of Techology,Changzhou 213001,China)

机构地区：[1]江苏理工学院汽车与交通工程学院,江苏常州213001

出　　处：《应用科技》2024年第2期120-126,共7页Applied Science and Technology

基　　金：江苏省研究生科研与实践创新计划项目(XSJCX22_48,XSJCX23_23).

摘　　要：针对多无人机(unmanned aerial vehicles,UAVs)协同控制问题,提出一种用于无人机协同编队控制方法。以虚拟结构控制为框架,结合梯度策略方法设计奖励函数和惩罚函数,将其看成一种引力和斥力运用到无人机协同编队控制系统,解决在多虚拟结构控制点下因编队数量增加导致机群的稳定性和协调性下降的问题。本文算法使得无人机编队既能协同运动,又能在运动中维持稳定的队形变换,还能防止机群之间的两机碰撞。对有控制输入的无人机编队系统进行仿真,验证了该方法的有效性。This paper proposes a method for cooperative formation control of multiple unmanned aerial vehicles(UAVs).It takes virtual structure control as the framework and combines the gradient strategy method to design the reward function and penalty function,which are regarded as a kind of gravitational force and repulsive force and applied to the UAVs cooperative formation control system to solve the problem of reduced stability and coordination of the UAVs cluster due to an increase in the number of formations under multiple virtual structure control points.This method not only ensures collaborative movement of UAVs formation,but also maintains stable formation transformation during the movement while preventing collisions between members within the cluster.The effectiveness of this method was verified through simulation of UAVs formation system with control inputs.

关 键 词：多无人机 协同控制 编队 虚拟结构算法 奖励函数 惩罚函数 碰撞 队形变换 

分 类 号：V21[航空宇航科学与技术—航空宇航推进理论与工程] V249