基于物理信息神经网络的飞机姿态预测模型研究

何森朋; 杨哲; 张玉刚; 杨文青

doi:10.7527/S1000-6893.2025.31850

航空学报 >

0 1 - 0

DOI: https://doi.org/10.7527/S1000-6893.2025.31850

基于物理信息神经网络的飞机姿态预测模型研究

何森朋 ,
杨哲 ,
张玉刚 ,
杨文青

展开

1. 西北工业大学航空学院
2. 中国飞行试验研究院
3. 飞行器基础布局全国重点实验室
4. 西北工业大学

收稿日期: 2025-01-27

修回日期: 2025-06-26

网络出版日期: 2025-06-27

基金资助

民用飞机典型特种试飞关键技术

收起

Research on aircraft attitude prediction model based on physics informed neural networks

HE Sen-Peng ,
YANG Zhe ,
ZHANG Yu-Gang ,
YANG Wen-Qing

Expand

Received date: 2025-01-27

Revised date: 2025-06-26

Online published: 2025-06-27

Supported by

the Key Technologies of Typical Special Test Flight of Civil Aircraft

Fold

摘要

为解决飞机试飞过程中由于超出飞行包线而导致的试飞安全事故问题，降低事故风险，根据飞机当前状态信息和舵面偏转的角度数据，预测飞机后续姿态演化过程，为辅助飞行员决策提供依据。结合飞行动力学方程与物理信息神经网络方法（PINNs），构建了飞机姿态实时预测模型（FD-PINNs模型），解决了基于神经网络的飞机姿态预测模型（NN模型）预测精度不足、泛化能力差的问题。通过FlightGear获取飞行仿真数据对模型进行了验证，计算结果表明FD-PINNs模型比NN模型泛化能力更强，预测精度更高，其中迎角预测结果的均方误差降低了68.5%。

关键词： 姿态预测; 物理信息神经网络; 飞行动力学; 试飞安全; 飞行试验

本文引用格式

何森朋 , 杨哲 , 张玉刚 , 杨文青 . 基于物理信息神经网络的飞机姿态预测模型研究[J]. 航空学报, 0 : 1 -0 . DOI: 10.7527/S1000-6893.2025.31850

Abstract

In order to solve the flight test safety accidents caused by exceeding the flight envelope during flight test, and reduce the accident risk, the subsequent attitude evolution process of the aircraft was predicted according to the current state information of the aircraft and the Angle data of rudder deflection, so as to provide a basis for auxiliary pilot decision-making. Combining flight dynamics equation and physical information neural network (PINNs), a real-time aircraft attitude prediction model (FD-PINNs model) was constructed to solve the problem of insufficient prediction accuracy and poor generalization ability of the neural net work-based aircraft attitude prediction model (NN model). FlightGear obtained flight simulation data to verify the model. The calculation results show that FD-PINNs model has stronger generalization ability and higher prediction accuracy than NN model, and the mean square error of Angle of attack prediction results is reduced by 68.5%.

Key words： attitude prediction; physical information neural networks; flight dynamics; flight test safety; flight test

参考文献

[1]WANG G Z, PEI B B, XU H J, et al. Risk quantifica-tion and visualization method for loss-of-control sce-narios in flight[J]. Aerospace, 2023, 10(5): 416. DOI:10.3390/aerospace10050416.
[2]WANG G Z, XU H J, PEI B B. An intelligent approach for flight risk prediction under icing conditions[J]. Chinese Journal of Aeronautics, 2023, 36(6):109-127. DOI:10.1016/j.cja.2023.02.020.
[3]KANG C, WOOLSEY C A. Model-based path predic-tion for fixed-wing unmanned aircraft using pose esti-mates[J]. Aerospace Science and Technology, 2020, 105: 106030.
[4]袁修久, 赵学军, 李嘉林. 基于航迹的飞机姿态角建模与仿真[J]. 系统工程与电子技术, 2016, 38(04): 889-894.YUAN X J, ZHAO X J, LI J L. Modeling and simula-tion of aircraft attitude angle based on air-path[J]. Sys-tems Engineering and Electronics, 2016, 38(4): 889-894(in Chinese).
[5]孙博, 高永卫, 魏斌斌. 无人机离机轨迹与姿态的高准度快速预测方法研究[J]. 弹箭与制导学报, 2023, 43(06): 97-104.SUN B, GAO Y W, WEI B B. Research on a fast and precise prediction method for UAV separation trajecto-ry and attitude[J]. Journal of Projectiles, Rockets, Mis-siles and Guidance, 2023, 43(06): 97-104(in Chinese).
[6]李海泉, 陈小前, 张嘉图, 等. 随机激励作用下飞机飞行姿态动力学研究[J]. 航空学报, 2022, 43(03): 225232.LI H Q, CHEN X Q, ZHANG J T, et al. Study on air-craft attitude dynamics under random excitation[J]. Ac-ta Aeronautica et Astronautica Sinica, 2022, 43 (3): 225232 (in Chinese). DOI: 10.7527/S1000-6893.2021.25232.
[7]杨丽梅, 郭立红. 一种序列图像飞行器姿态的预测方法[J]. 计算机测量与控制, 2006, 14(06): 797-799.
YANG L M, GUO L H. Aircraft pose prediction meth-od of image sequences[J]. Computer Measurement & Control, 2006, 14(06): 797-799(in Chinese).
[8]PASHA A A, SANKARALINGAM L, RAHMAN M M, et al. MEMS fault-tolerant machine learning algo-rithm assisted attitude estimation for fixed-wing UAVs[J]. Engineering Applications of Artificial Intelli-gence, 2024, 129: 107608. DOI:10.1016/j.engappai.2023.107608.
[9]曹咏弘, 张慧, 马铁华, 等. 基于神经网络的无陀螺捷联惯导系统姿态预测[J]. 中国惯性技术学报, 2008, 16(02): 159-161+170.CAO Y H, ZHANG H, MA T H, et al. Attitude forecast of gyroscope-free SINS based on neural network[J]. Journal of Chinese Inertial Technology, 2008, 16(02): 159-161+170(in Chinese).
[10]梁少军, 郑幸, 谢礼鹏, 等. 固定翼无人机纵向控制回路多工况状态预测[J]. 兵器装备工程学报, 2020, 41(10): 203-209.LIANG S J, ZHENG X, XIE L P, et al. Muti-Condition state prediction of longitudinal control loop of fixed-wing UAV[J]. Journal of Ordnance Equipment Engi-neering, 2020, 41(10): 203-209(in Chinese).
[11]路晶, 史宇, 任洲. 基于增量式在线学习飞行训练姿态预测研究[J]. 航空计算技术, 2022, 52(5): 5-8.LU J, SHI Y, REN Z. Research on Attitude Prediction of Flight Training Based on Incremental Online Learn-ing [J]. Aeronautical Computing Technique, 2022, 52(5): 5-8(in Chinese).
[12]RAISSI M, PERDIKARIS P, KARNIADAKIS G E. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. Journal of Computational Physics, 2019, 378: 686-707. DOI: 10.1016/j.jcp.2018.10.045.
[13]CAO W B, SONG J H, ZHANG W W. A solver for subsonic flow around airfoils based on physics-informed neural networks and mesh transformation[J]. Physics of Fluids, 2024, 36(2):027134. DOI: 10.1063/5.0188665.
[14]李野, 陈松灿. 基于物理信息的神经网络: 最新进展与展望[J]. 计算机科学, 2022, 49(4): 254?262.LI Y, CHEN S C. Physics-informed Neural Networks: Recent Advances and Prospects[J]. Computer Science, 2022, 49(4): 254?262(in Chinese).
[15]PANG G F, LU L, KARNIADAKIS G E. FPinns: Fractional physics informed neural networks[J]. SIAM Journal on Scientific Computing, 2019, 41(4): A2603-A2626. DOI:10.1137/18M1229845.
[16]HE G Y, ZHAO Y X, YAN C L. MFLP-PINN: A phys-ics-informed neural network for multiaxial fatigue life prediction[J]. European Journal of Mechanics - A/Solids, 2023,98:104889. DOI: 10.1016/j.euromechsol.2022.104889.
[17]BAI Z W, SONG S F. Physics-informed neural network for first-passage reliability assessment of structural dy-namic systems[J]. Computers & Structures, 2023,289:107189. DOI: 10.1016/j.compstruc.2023.107189.
[18]MOHAJERIN N, MOZIFIAN M, WASLANDER S. Deep learning a quadrotor dynamic model for multi-step prediction[C]. 2018 IEEE International Confer-ence on Robotics and Automation (ICRA), 2018, 2454-2459.
[19]GU W, PRIMATESTA S, RIZZO A. Physics-informed neural network for quadrotor dynamical modeling[J]. Robotics And Autonomous Systems, 2024, 171, 104569.
[20]王康晋. 基于PINN的细长体飞行器气动特性预测研究[D]. 绵阳: 西南科技大学, 2024.WANG K J, Research on predicting aerodynamic char-acteristics of slender aircraft based on PINN[D]. Mian yang: Southwest University of Science and Technology, 2024: 13-15 (in Chinese).
[21]付军泉, 钟伯文, 钟运琴, 等. 基于物理信息神经网络的飞机气动参数辨识方法[J]. 空气动力学学报, 2023, 41(09): 30-37.FU J Q, ZHONG B W, ZHONG Y Q, et al. A physics informed neural network based method for aircraft aer-odynamic parameter identification[J]. Acta Aerody-namica Sinica, 2023, 41(9): 30-37(in Chinese).
[22]方振平, 陈万春, 张曙光. 航空飞行器飞行动力学[M]. 北京: 北京航空航天大学出版社, 2005: 22-24, 174-181.FANG Z P, CHEN W C, ZHANG S G. Aerodynamic flight dynamics of aircraft[M]. Beijing: Beihang Uni-versity Press, 2005: 22-24, 174-181.
[23]金长江. 飞行动力学：飞机飞行性能计算[M]. 北京: 国防工业出版社, 1983: 29-34.JIN C J. Flight dynamics: aircraft flight performance calculation[M]. Beijing: National Defense Industry Press, 1983: 29-34.

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献