基于数据滤波的带协方差重置的递推贝叶斯算法
针对传统最小二乘算法计算量大、在有色噪声干扰下估计有误差的问题,提出了一种基于滤波技术的带协方差重置的递推贝叶斯算法。该算法使用一个动态非线性滤波器对输入输出数据进行滤波,然后使用贝叶斯方法进行参数估计。为了加快参数的收敛速度,在算法中加入了一种新型的协方差重置策略。计算量分析表明,当过程模型和噪声模型的阶数分别为6和4的时候,所提算法可以减少约62.35%的计算量。仿真结果显示,所提算法与传统最小二乘算法在采样数据长度为3 000时的估计误差分别为0.771%和1.118%。因此,所提算法具有较高的计算效率,并且可以给出精度较高的参数估计值。...
Saved in:
| Published in | 计算机应用研究 Vol. 33; no. 5; pp. 1338 - 1341 |
|---|---|
| Main Author | |
| Format | Journal Article |
| Language | Chinese |
| Published |
淮安信息职业技术学院 电气工程系,江苏 淮安 223003
2016
江苏大学 电气信息工程学院,江苏 镇江212013%江苏大学 电气信息工程学院,江苏 镇江,212013 |
| Subjects | |
| Online Access | Get full text |
| ISSN | 1001-3695 |
| DOI | 10.3969/j.issn.1001-3695.2016.05.013 |
Cover
| Summary: | 针对传统最小二乘算法计算量大、在有色噪声干扰下估计有误差的问题,提出了一种基于滤波技术的带协方差重置的递推贝叶斯算法。该算法使用一个动态非线性滤波器对输入输出数据进行滤波,然后使用贝叶斯方法进行参数估计。为了加快参数的收敛速度,在算法中加入了一种新型的协方差重置策略。计算量分析表明,当过程模型和噪声模型的阶数分别为6和4的时候,所提算法可以减少约62.35%的计算量。仿真结果显示,所提算法与传统最小二乘算法在采样数据长度为3 000时的估计误差分别为0.771%和1.118%。因此,所提算法具有较高的计算效率,并且可以给出精度较高的参数估计值。 |
|---|---|
| Bibliography: | 51-1196/TP Traditional least squares identification algorithm required much computational cost and its estimates were biased when the noise was colored. To overcome these shortcomings,this paper proposed a filter based recursive Bayesian identification algorithm with covariance resetting. In this algorithm,it firstly filtered the input and output data by a dynamics nonlinear filter and then used recursive Bayesian algorithm to estimate parameters. It also integrated a modified covariance resetting method to the algorithm. Analysis revealed that the proposed algorithm could reduce the computational burden by 62. 35% compared with recursive Bayesian algorithm. Simulations indicate that the estimation errors of the two algorithms are 0. 771% and 1. 118% respectively. So the proposed algorithm has higher efficiency and can generate estimates with higher accuracy. recursive Bayesian algorithm; filter; covariance resetting; parameter estimation; on-line algorithm; pseudolinear model Jing Shaoxue;Li Zhengming( 1. Dept. |
| ISSN: | 1001-3695 |
| DOI: | 10.3969/j.issn.1001-3695.2016.05.013 |