摘要
针对实际飞行中无法直接测量的扑翼飞行器气动参数辨识问题,结合刚体六自由度模型,提出一种基于迭代学习和人工鱼群法的扑翼飞行器隐式气动参数辨识方法。鉴于扑翼飞行器飞行试验中待辨识气动参数值难以直接测量、导致一般辨识算法中梯度难以求解的问题,提出基于摄动法的梯度方向寻优方法。考虑到待辨识参数数量及辨识结果对参数初始值的敏感性,该方法采用人工鱼群算法优化计算待辨识参数初始值。且针对迭代过程中损失函数易陷入局部最优和优化速度受限问题,采用变学习因子迭代学习策略。试验结果表明,所提出的算法能有效估算出扑翼飞行器气动参数。
To identify the aerodynamic parameters of flapping-wing micro aerial vehicle(FMAV) that cannot be directly measured in actual flight, an implicit aerodynamic parameter identification method based on iterative learning and artificial fish swarm method is proposed with the rigid body six-degree-of-freedom model. Considering the difficulty in calculating the gradient with the general identification algorithm, a gradient direction optimization method based on perturbation method is proposed. In view of the number of parameters to be identified and the sensitivity of the identification results to the initial parameters values, an artificial fish swarm algorithm is applied to optimize the initial parameters values. To solve the problems that the cost is easy to fall into the local optimum and the optimization speed is limited in the iterative process, the variable learning factor iterative learning strategy is adopted. Experimental tests show that the proposed algorithm can effectively estimate the aerodynamic parameters of FMAV based on the measured flight data.
作者
孙逊莱
杜昌平
叶志贤
陈丹
王柯钦
郑耀
SUN Xun-lai;DU Chang-ping;YE Zhi-xian;CHEN Dan;WANG Ke-qin;ZHENG Yao(School of Aeronautics and Astronautics,Zhejiang University,Hangzhou 310027,China)
出处
《飞行力学》
CSCD
北大核心
2019年第6期17-21,共5页
Flight Dynamics
基金
装备预研教育部联合基金(重点)项目(6141A02011803)
关键词
迭代学习
非线性
气动参数
摄动法
扑翼
iterative learning
nonlinear
aerodynamic parameter
perturbation method
flapping wing
