改进Quatre算法的无人机编队快速集结方法  

UAV formation rapid assembly method based on improved Quatre algorithm

作　　者：孙田野 孙伟 吴建军[2] SUN Tianye;SUN Wei;WU Jianjun(School of Aerospace Science and Technology,Xidian University,Xi’an 710071,China;The 365th Research Institute,Northwestern Polytechnical University,Xi’an 710129,China)

机构地区：[1]西安电子科技大学空间科学与技术学院,陕西西安710071 [2]西北工业大学第365研究所,陕西西安710129

出　　处：《系统工程与电子技术》2022年第9期2840-2848,共9页Systems Engineering and Electronics

基　　金：国家自然科学基金(61671356);陕西省重点研发计划(2020GY-136);陕西省重点研发计划重点产业创新链项目(2019ZDLGY14-02-03,2022ZDLGY03-01);中国高校产学研创新基金(2021ZYA08004);西安市科技计划(2022JH-RGZN-0039)资助课题。

摘　　要：为了提高多无人机在编队集结过程中的稳定性,提出基于改进拟仿射类进化(improved quasi-affine transformation evolutionary,IQuatre)算法的无人机编队集结方法。首先,基于分布式模型预测控制(distributed model predictive control,DMPC)建立无人机编队的运动预测模型,通过预测无人机的“未来态”规避编队内碰撞风险,滚动优化的数学模型提高了无人机到达指定位置的稳定性,使得无人机更好、更快地加入编队飞行;其次,对Quatre算法进行种群优化改进,将携带最优基因的父代个体有选择性地加入子代种群,加快种群收敛。实验结果表明,基于DMPC的无人机编队集结未出现碰撞情况,减小了无人机调整状态过程中出现的位置偏差;对比仿真验证了IQuatre算法能够提高编队集结的稳定性,较原Quatre算法减少了5.2%的平均迭代次数,在计算时间上节约了4.6%,位置误差减小了0.45 m。To improve the stability of multiple unmanned aerial vehicle(UAV)in the process of formation assembly,this paper proposes a UAV formation method based on improved quasi-affine transformation evolutionary(IQuatre)algorithm.First of all,based on distributed model predictive control(DMPC),a motion prediction model of the UAV formation is established.By predicting the“future state”of the UAV,the risk of collision within the formation is avoided.The mathematical model of rolling optimization improves the stability of the UAV to reach the designated position.UAV is better and faster to join the formation flight.Secondly,the Quatre algorithm is improved on population optimization by the parent individuals carrying the best genes selectively added to the offspring population to speed up population convergence.The experimental results show that the UAV formation based on the DMPC has no UAV collision,and reduces the position deviation during the UAV adjustment state.Comparing simulations verify that the IQuatre algorithm can improve the stability of the formation assembly.Compared with the original Quatre algorithm,the average number of iterations is reduced by 5.2%,the calculation time is saved by 4.6%,and the position error is reduced by 0.45 m.

关 键 词：无人机编队 分布式模型预测控制 改进拟仿射类进化算法 滚动优化 种群优化 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]