摘要
对再入式高超声速飞行器的气动参数在线辨识方法进行了分析研究,采用滤波器对动态方程进行静态化处理,以简化辨识方法,但同时引入了不确定的滤波器参数.为了减小辨识过程中由滤波器参数选择引起的辨识误差,设计了一种参数选择策略.在常规选择参数的基础上引入了智能优化算法——粒子群优化算法,用以确定合适的滤波器参数值.然后,利用基于带遗忘因子的最小二乘法对时变气动参数进行在线辨识.最后基于SX-2模型进行了相关仿真.结果表明:基于粒子群优化算法的气动参数在线辨识方法与未引入参数选择策略的气动参数在线辨识方法相比,辨识精度得到了一定程度的提高.
The aerodynamic parameters online identification methods for the reentry hypersonic vehicle were researched,and the filter was used to process dynamic equation statically so that identification methods could be simplified,but the inappropriate filter parameters were introduced at the same time.A parameter selection strategy was designed to reduce the identification error,which caused by the adoption of filter parameters during the identification procedure.On the basis of conventional parameter selection methods,particle swarm optimization（PSO）was proposed to determine a set of more appropriate filter parameters during the identification.Then,the time-varying aerodynamic parameters were identified online by the least squares method with forgetting factor.In the end,based on example model SX-2,the simulation was conducted.The results show that the method based on PSO ensures higher recognition accuracy than the approaches without parameters selection strategy.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2016年第3期116-120,共5页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(61473124
61203081
61174079)
教育部博士学科点专项科研基金资助项目(20120142120091)
科技部国际合作项目(2012DFG70640)
关键词
粒子群优化
频谱分析
参数辨识
最小二乘算法
滤波器
在线
particle swarm optimization
spectrum analysis
parameter identification
least squares approximations
filters
online
