摘要
在噪声未知但有界的情况下,为了解决辨识模型回归矩阵奇异条件下的参数集员辨识问题,基于差分进化算法和非线性向量回归,提出一种新颖的四旋翼参数集员辨识方法。利用非线性向量回归方法建立逼近参数向量与误差向量的范数之间的函数关系模型,并用差分进化算法优化模型参数,避免参数选择的盲目性。仿真结果验证了这种方法的有效性。
Assuming that noise is unknown but bounded, in order to solve parameter set membership iden-tification problem under condition that regression matrix of identification scheme is ill-conditioned. A no- vel set membership identification method was proposed based on differential evolution (DE) and nonlin- ear vector regression. A nonlinear vector regression method was applied to build a model which approxi- mated the functional relationship between the parameter vector and the norms of the error vector. DE al- gorithm is used to optimize the model parameters which avoid the arbitrary. The simulating results vali- date the effectiveness of this method.
出处
《飞行力学》
CSCD
北大核心
2015年第4期334-338,348,共6页
Flight Dynamics
基金
高等学校博士学科点专项科研基金资助(20121102110008)
关键词
无人飞行器
差分进化算法
微分器
向量回归
集员辨识
quadrotor UAV
differential evolution
differentiator
vector regression
set membership iden-tification