摘要
提出一种基于输入输出信号的三阶累积量 (third ordercumulants)对非高斯AR模型阶次p进行辨识的新算法 .该算法依模型阶次递推 ,通过求解代价函数J(^p) 的首次最小值 ,确定AR模型的阶次估计值 ^p .理论分析和仿真结果均显示 ,该算法具有良好的收敛性和精确性 .
An order recursive algorithm for determining the order \$p\$ of non Gaussian AR models based on the third order cumulants is suggested in this paper. The order estimates of AR models can be selected by achieving the first minimum point of a well defined cost function J().Theoretical analyses and simulation results illustrate that the proposed algorthm, which has a strong convergence behaviour,is accurate and simple.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2000年第2期281-285,共5页
Control Theory & Applications
基金
广东省自然科学基金!(960 30 4 )
关键词
随机系统
AR模型
三阶累积量
系统辨识
阶次辨识
stochastic systems
AR models
third order cumulants
system order identification
