随机噪声框架下非伪频率响应的计算方法  

Computation of Plant Unfalsified Frequency Response with a Prescribed Probability

作　　者：朱凌[1] 

机构地区：[1]清华大学自动化系,北京100084

出　　处：《控制与决策》2005年第9期1022-1025,共4页Control and Decision

基　　金：国家自然科学基金项目(60174022)

摘　　要：推导了当系统输入输出实验数据精确已知时,由未建模动态产生的频率响应集合解析表达式.采用序贯蒙特卡洛方法产生随机噪声序列,由这些噪声序列计算出相应的非伪频率响应集.提出一个找到包含所有频率响应且有最小半径的圆的数值优化算法,通过该算法,得到一个以指定概率包含控制对象非伪频率响应的最小集合.最后给出了仿真实例,验证了整个算法的有效性.With exact input and output experimental data, an analytical expression is derived for the unfalsified frequency response consistent with prior knowledge and the experimental data. By using a sequential Monte Carlo method, noise series are sampled and the corresponding unfalsified frequency response sets are computed. A numerical algorithm is proposed to find the circle with minimal radius and containing all the given circles. With this algorithm, a set is obtained which includes the plant actual frequency response with a prescribed probability. Finally, a numerical simulation shows the validity of this method.

关 键 词：鲁棒辨识 非伪频率响应 随机实验 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]