摘要
针对固定时间下的两航天器三维空间追逃问题,采用协同进化算法将复杂的双边最优规划问题简化成对追逃过程中纳什均衡点的搜索,进而得出追逃双方的最优对抗策略及解算方法。考虑在保证算法计算精度的前提下缩短计算时间,将对策模型进行简化处理,以航天器推力指向角为控制量对协同进化算法进行编码设计,利用B样条基函数对编码进行逼近拟合。两航天器均为连续小推力作用,以二者的末端相对距离作为支付函数,逃逸器希望支付最大,追踪器希望支付最小,并依此建立共享适应度函数。所提方法中对协同算子进行改进设计,提高算法空间搜索能力,并采用精英保留策略提高算法的收敛速度。仿真算例得到追逃双方的最优控制策略及相应的追逃轨迹,表明所提方法的能够解决此类航天器追逃问题。
The co-evolutionary algorithm is used to simplify the complex double boundary optimal planning problems into the searching of Nash equilibrium in three dimensional space pursuit-evasion of two spacecraft within a fixed time interval.This method provides an optimal antagonistic strategy and its solution for the pursuer and evader spacecraft.To reduce the computing time while maintaining computational accuracy,the strategy model is simplified with spacecraft thruster directional angles approached using the B-spline basis function as control variables.The relative terminal distance of the two spacecraft,which are both considered to below continuous thrust systems is defined as a pay off function which the evader expects to be maximum while the pursuer expects to be minimum.A shared fitness function is therefore established.This study also improves the design of compact genetic operators which enhance the space searching ability of the algorithm.The use of the elitism strategy in the algorithm enables a faster converging rate.Simulation results show that the optimal control strategy and corresponding pursuit and evasion trajectories for both spacecraft are obtained by the proposed method,thus the ability of the algorithm in solving two spacecraft pursuit-evasion problems is demonstrated.
作者
刘源
李玉玲
郝勇
周俊峰
LIU Yuan;LI Yuling;HAO Yong;ZHOU Junfeng(College of Automation, Harbin Engineering University, Harbin 150001, China)
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2018年第4期868-877,共10页
Systems Engineering and Electronics
基金
国家自然科学基金(61273081)
黑龙江省博士后科研启动基金(LBH-Q14054)
中央高校基本科研业务费专项资金(HEUCFD1503)资助课题
关键词
协同进化
空间追逃
纳什均衡
B样条拟合
精英保留
compact genetic
space pursuit-evasion
Nash equilibrium
B-spline fitting
elitism preservation
