基于WOA-DNN的高超声速飞行器实时再入轨迹优化方法

Real-time Reentry Trajectory Optimization Method for Hypersonic Vehicles Based on WOA-DNN

摘要

针对高超声速飞行器再入轨迹优化的实时性需求,提出了一种基于鲸鱼优化算法(whale optimization algorithm,WOA)与深度神经网络(deep neural network,DNN)结合的实时轨迹优化方法。首先,建立高超声速飞行器再入轨迹优化模型,采用序列二阶锥规划方法,将原本的非凸最优控制问题转化为凸优化问题,求解生成包含气动参数不确定性的最优轨迹数据集。其次,构建以飞行器状态序列为输入、最优倾侧角指令为输出的DNN模型。针对DNN的性能对其初始权重与阈值等超参数敏感度高、依赖性强的问题,引入WOA对上述参数进行全局优化搜索,旨在显著提升模型的预测精度与泛化能力。在线规划阶段,训练完成的网络能够根据当前飞行状态实时生成近似最优的控制指令。数值仿真结果表明,在标称及气动不确定条件下,所提WOA-DNN轨迹优化方法能够快速生成满足终端精度要求的可行轨迹,显著提升计算效率,充分展现了其在精度与鲁棒性上的综合优势。

Abstract

To address the real-time requirements for hypersonic vehicle reentry trajectory optimization, a real-time trajectory optimization method that integrates the whale optimization algorithm (WOA) with deep neural network (DNN) was proposed. Firstly, a reentry trajectory optimization model for a hypersonic vehicle was established. The original non-convex optimal control problem was transformed into a convex optimization problem for efficient solution via sequential second-order cone programming, generating an optimal trajectory dataset incorporating aerodynamic parameter uncertainties. Subsequently, a DNN was constructed, mapping the vehicle's state sequence to optimal bank-angle commands. To address the high sensitivity of DNN performance to hyperparameters such as initial weights and thresholds, the WOA was introduced to globally optimize these parameters, thereby significantly enhancing the prediction accuracy and generalization capability. In the final online planning stage, near-optimal control commands were generated in real time based on the actual flight states. Numerical simulations demonstrate that under nominal and aerodynamic uncertainty conditions, the proposed WOA-DNN optimization method rapidly generates feasible trajectories that satisfy terminal accuracy requirements, significantly enhances computational efficiency. This highlights comprehensive advantages of the method in terms of both precision and robustness for trajectory optimization.

关键词

高超声速飞行器 / 再入轨迹优化 / 深度神经网络 / 鲸鱼优化算法 / 序列二阶凸规划

Key words

hypersonic vehicle / reentry trajectory optimization / deep neural networks / whale optimization algorithm / sequential second-order cone programming

参考文献

[1] 王鹏飞, 刘红琴, 杨自牧, 等. 吸气式高超声速飞行器控制研究进展[J]. 空天技术, 2024(5): 33-43.

[2] WANG Pengfei, LIU Hongqin, YANG Zimu, et al. Review of control on air-breathing hypersonic flight vehicle[J]. Aerospace Technology, 2024(5): 33-43. (in Chinese)

[3] LIU S, YAN B, HUANG W, et al. Current status and prospects of terminal guidance laws for intercepting hypersonic vehicles in near space: a review[J]. Journal of Zhejiang University-SCIENCE A, 2023, 24(5): 387-403.

[4] 魏昊, 蔡光斌, 凡永华, 等. 高超声速飞行器再入滑翔段在线制导[J]. 北京航空航天大学学报, 2025, 51(1): 183-192.

[5] WEI Hao, CAI Guangbin, FAN Yonghua, et al. Online guidance for hypersonic vehicles in glide-reentry segment[J]. Journal of Beijing University of Aeronautics and Astronautics, 2025, 51(1): 183-192. (in Chinese)

[6] 徐慧, 蔡光斌, 崔亚龙, 等. 高超声速滑翔飞行器再入轨迹优化[J]. 哈尔滨工业大学学报, 2023, 55(4): 44-55.

[7] XU Hui, CAI Guangbin, Cui Yalong, et al. Reentry trajectory optimization method of hypersonic glide vehicle[J]. Journal of Harbin Institute of Technology, 2023, 55(4): 44-55. (in Chinese)

[8] 杨明, 刘明, 葛亚杰, 等. 基于准平衡滑翔的再入轨迹解析规划方法[J]. 弹道学报, 2020, 32(4): 20-26.

[9] YANG Ming, LIU Ming, GE Yajie, et al. Analytical method of planning reentry trajectory via quasi-equilibrium gliding condition[J]. Journal of Ballistics, 2020, 32(4): 20-26. (in Chinese)

[10] QING W, PAN B. Modified fractional homotopy method for solving nonlinear optimal control problems[J]. Nonlinear Dynamics, 2024, 112: 3453-3479.

[11] SARANATHAN H, GRANT M J. Incorporation of rigid body dynamics into indirect hypersonic trajectory optimization[J]. Journal of Spacecraft and Rockets, 2024, 61(2): 393-406.

[12] SIDHOUMY , OGURI K. On the performance of different smoothing methods for indirect low-thrust trajectory optimization[J]. The Journal of the Astronautical Sciences, 2023, 70(6): 51.

[13] ZHANG Y, ZHU B, CHENG M, et al. Trajectory optimization for spacecraft autonomous rendezvous and docking with compound state-triggered constraints[J]. Aerospace Science and Technology, 2022, 127: 107733.

[14] ZHANG G, WEN C, QIAO D, et al. Trajectory planning for satellite swarms with nonlinear terminal constraints using penalty concave relaxation[J]. Aerospace Science and Technology, 2024, 144: 108826.

[15] MORELLI A, MORESELLI A, GIORDANO C, et al. Convex trajectory optimization using thrust regularization[J]. Journal of Guidance, Control, and Dynamics, 2024, 47(2): 339-346.

[16] 薛光伟, 辛万青, 傅瑜. 能量优化分配再入轨迹快速规划方法[J]. 弹道学报, 2023, 35(2): 1-8.

[17] XUE Guangwei, XIN Wanqing, FU Yu. Reentry trajectory rapid planning algorithm based on energy allocation optimization[J]. Journal of Ballistics, 2023, 35(2): 1-8. (in Chinese)

[18] 闫循良, 王培臣, 王舒眉, 等. 基于混沌多项式的RBCC飞行器上升段鲁棒轨迹快速优化[J]. 航空学报, 2023, 44(21): 528349.

[19] YAN Xunliang, WANG Peicheng, WANG Shumei, et al. Rapid robust trajectory optimization for RBCC vehicle ascent based on polynomial chaos[J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(21): 528349. (in Chinese)

[20] 伍元培, 王阳, 赵爱红, 等. 基于深度学习网络的高超声速飞行器避障轨迹规划方法[J]. 战术导弹技术, 2022(1): 58-64.

[21] WU Yuanpei, WANG Yang, ZHAO Aihong, et al. Obstacle avoidance trajectory planning of hypersonic vehicle based on deep learning network[J]. Tactical Missile Technology, 2022(1): 58-64. (in Chinese)

[22] CHAI R, TSOURDOS A, SAVVARIS A, et al. Real-time reentry trajectory planning of hypersonic vehicles: a two-step strategy incorporating fuzzy multiobjective transcription and deep neural network[J]. IEEE Transactions on Industrial Electronics, 2019, 67(8): 6904-6915.

[23] CHAI R, TSOURDOS A, SAVVARIS A, et al. Six-DOF spacecraft optimal trajectory planning and real-time attitude control: a deep neural network-based approach[J]. IEEE Transactions on Neural Networks and Learning Systems, 2019, 31(11): 5005-5013.

[24] YIN S, LI J, CHENG L. Low-thrust spacecraft trajectory optimization via a DNN-based method[J]. Advances in Space Research, 2020, 66(7): 1635-1646.

[25] SHI Y, WANG Z. Onboard generation of optimal trajectories for hypersonic vehicles using deep learning[J]. Journal of Spacecraft and Rockets, 2021, 58(2): 400-414.

[26] WANG J, WU Y, LIU M, et al. A real-time trajectory optimization method for hypersonic vehicles based on a deep neural network[J]. Aerospace, 2022, 9(4): 188.

[27] MIRJALILI S, LEWIS A. The whale optimization algorithm[J]. Advances in Engineering Software, 2016, 95: 51-67.

[28] DAI P, FENG D, FENG W, et al. Entry trajectory optimization for hypersonic vehicles based on convex programming and neural network[J]. Aerospace Science and Technology, 2023, 137: 108259.

[29] 徐慧, 蔡光斌, 穆朝絮, 等. 高超声速滑翔飞行器全程总红外辐射最小的轨迹优化(特邀)[J]. 红外与激光工程, 2022, 51(4): 61-73.

[30] XU Hui, CAI Guangbin, MU Chaoxu, et al. Trajectory optimization of hypersonic glide vehicle with minimum total infrared radiation (invited)[J]. Infrared and Laser Engineering, 2022, 51(4): 61-73. (in Chinese)

基金资助

国家自然科学基金项目(62473374)

国家自然科学基金项目(62403487)