基于随机响应面与混沌多项式的鲁棒轨迹优化  

Robust trajectory optimization method based on stochastic response surface and polynomial chaos

作　　者：王培臣 闫循良[1] 王宽 郑雄[2] WANG Peichen;YAN Xunliang;WANG Kuan;ZHENG Xiong(School of Astronautics,Northwestern Polytechnical University,Xi’an 710072,China;China Academy of Launch Vehicle Technology,Beijing 100076,China)

机构地区：[1]西北工业大学航天学院,陕西西安710072 [2]中国运载火箭技术研究院,北京100076

出　　处：《系统工程与电子技术》2023年第10期3226-3239,共14页Systems Engineering and Electronics

基　　金：国家自然科学基金(11602296);陕西省自然科学基础研究计划(2019JM-434);中央高校基本科研业务费专项资金(G2022KY0613)资助课题。

摘　　要：为提升多参数不确定性条件下鲁棒轨迹优化求解技术的通用性和效率,设计了一种基于随机响应面法与混沌多项式的鲁棒轨迹优化方法。首先,将非嵌入式混沌多项式与随机响应面法相结合,引入拉丁超立方采样策略,构建了具有普适性的多维不确定性量化传播模型,在实现不确定随机问题有效转化的同时,保证了问题转化的精度和计算效率;随后,将该模型与凸优化方法相结合,设计了一种基于序列凸优化算法的确定性轨迹优化求解策略,以实现对该高维确定性优化问题的快速求解;最终,以考虑初始状态和过程参数等多参数不确定性的高超声速滑翔轨迹优化问题为例进行了仿真,并与现有方法进行了对比分析。结果表明,与传统确定性轨迹优化算法相比,所提方法可获得兼具可靠性和鲁棒性的高超声速滑翔优化轨迹;与现有典型鲁棒轨迹优化方法相比,所提方法具有相当的精度和显著的计算效率优势,且在不确定性维数增多的情况下,仍能保证鲁棒轨迹优化生成的高效性和可靠性。To improve the generality and computational efficiency of robust trajectory optimization technology under multi parameter uncertainty,a robust trajectory optimization method based on stochastic response surface method and polynomial chaos is designed.Firstly,non-intrusive polynomial chaos is combined with stochastic response surface,and a Latin hypercube sampling strategy is introduced to establish a multi-dimensional uncertainty quantification model,which realizes the effective and accurate transformation of the uncertain optimization problem.Secondly,combining the model with the convex optimization method,a trajectory optimization solution strategy based on the sequential convex optimization algorithm is designed to realize the rapid solution of the high-dimensional deterministic optimization problem.Finally,a hypersonic glide trajectory optimization problem considering multi parameter uncertainty such as initial state and process parameters is simulated and compared with the existing methods.The results show that compared with the traditional deterministic trajectory optimization algorithm,the proposed method can obtain the hypersonic glide trajectory with both reliability and robustness.Compared with the existing robust trajectory optimization methods,the proposed method has considerable accuracy and significant computational efficiency,which can ensure the efficiency and reliability of the robust trajectory optimization trajectory generation when the dimension of random variables increases.

关 键 词：不确定性 鲁棒轨迹优化 混沌多项式 随机响应面 凸优化 高超声速滑翔 

分 类 号：V412.44[航空宇航科学与技术—航空宇航推进理论与工程]