一种基于MISOCP的复杂绕飞区轨迹优化方法
An MISOCP Based Trajectory Optimization Method for Complex No-fly Zone
陈扬 1梁卓 1吕瑞 1喻涛 1张雪婷1
作者信息
- 1. 中国运载火箭技术研究院,北京 100076
- 折叠
摘要
针对复杂绕飞问题,为提升优化方法的全局寻优能力,提出一种基于混合整数二阶锥规划(MISOCP)的迭代求解算法.首先,定量分析了传统序列凸化方法处理绕飞区(NFZ)约束产生的误差;随后,为消除误差,利用整数变量结合自变量变换将绕飞区约束转化为混合整数线性约束;最后,将原问题转化为MISOCP问题,并设计了迭代算法进行求解.仿真表明,本方法全局寻优能力强,且求解时间仅为GPOPS的20%.对随机生成的复杂绕飞场景,本方法鲁棒性较好,且平均求解时间在10 s量级.
Abstract
In the context of complex evasion problems,to enhance the global optimization capability,an iterative algorithm based on mixed-integer second-order cone programming(MISOCP)is proposed.Firstly,the quantified analysis of the errors caused by traditional successive convexification methods on no-fly zone(NFZ)constraints is conducted.Subsequently,to eliminate this error,NFZ constraints are reconstructed into mixed-integer linear constraints using integer variables and variable transformation.Lastly,the original problem is transformed into a MISOCP problem,and an iterative algorithm is designed for solving it.Simulation results demonstrate that,given a variety of different initial guess trajectories,our method has strong global optimization capabilities,and the computation time is only 20%of that required by GPOPS.For randomly generated complex NFZ scenarios,the proposed method shows good robustness,and the overall computational time within the range of 10 seconds.
关键词
飞行器/轨迹优化/二阶锥规划/复杂绕飞区/路径决策Key words
Vehicle/Trajectory optimization/Second-order cone programming/Complex no-fly zone/Path planning引用本文复制引用
出版年
2024
NETL
NSTL
万方数据