一种不完全信息下递推辨识方法及收敛性分析
针对信号在网络环境下传输带来不完全信息使得在线参数辨识算法和收敛性困难的问题,不同于传统递推最小二乘方法,本文提出了一种不完全信息下递推辨识方法并分析其收敛性.首先运用伯努利分布刻画引起不完全信息的数据丢包特性,然后基于辅助模型方法补偿不完全信息并构造了新的数据信息矩阵,并运用矩阵正交变换性质对数据信息矩阵进行QR分解,推导了融合网络参数的递推辨识新算法,理论证明了在不完全信息下递推参数辨识算法的收敛性.最后仿真结果验证了所提方法的可行性和有效性....
Saved in:
| Published in | 自动化学报 Vol. 41; no. 8; pp. 1502 - 1515 |
|---|---|
| Main Author | |
| Format | Journal Article |
| Language | Chinese |
| Published |
上海大学机电工程与自动化学院上海市电站自动化技术重点实验室上海 200072
2015
|
| Subjects | |
| Online Access | Get full text |
| ISSN | 0254-4156 1874-1029 |
| DOI | 10.16383/j.aas.2015.c140766 |
Cover
| Summary: | 针对信号在网络环境下传输带来不完全信息使得在线参数辨识算法和收敛性困难的问题,不同于传统递推最小二乘方法,本文提出了一种不完全信息下递推辨识方法并分析其收敛性.首先运用伯努利分布刻画引起不完全信息的数据丢包特性,然后基于辅助模型方法补偿不完全信息并构造了新的数据信息矩阵,并运用矩阵正交变换性质对数据信息矩阵进行QR分解,推导了融合网络参数的递推辨识新算法,理论证明了在不完全信息下递推参数辨识算法的收敛性.最后仿真结果验证了所提方法的可行性和有效性. |
|---|---|
| Bibliography: | Data packet dropout, parameter estimation, recursive least squares, matrix QR decomposition, algorithm convergence Under the network environment, the uncomplete infromation causes an undesirable effect on the parameter identification and convergence. Unlike the traditional recursive least squares （RLS） algorithm, the paper proposes a novel online recursive identification method with uncomplete communication constraints. In this algorithm, the Bernoulli process is firstly employed to describe the character of data packet losses, and the uncomplete information is compensated by the auxiliary model strategy. The new data information matrix is then constructed, which is decomposed by QR decomposition and the intermediate matrix can be updated recursively. An new recursive least squares （RLS） algorithm under networks with random packet losses is then presented, and its convergence is analysed. Simulation confirms the feasibility and efficiency of the proposed method. 11-2109/TP DU Da-Jun, SHANG Li-Li, QI Bo, FEI Min- |
| ISSN: | 0254-4156 1874-1029 |
| DOI: | 10.16383/j.aas.2015.c140766 |