论文部分内容阅读
为克服应用Least Mean Square(LMS),Normalized LMS(NLMS)或Recursive Least Square(RLS)算法估计二阶Volterra滤波器系数时参数选择不当引起的问题,提出了基于后验误差假设并具有可变收敛因子的Davidon-Fletcher-Powell(DFP)方法的二阶Volterra自适应滤波器(DFPSOVF).给出参数估计算法中自相关逆矩阵估计的递归更新公式,并对算法的计算复杂度进行了分析.应用DFPSOVF滤波器对纯净和不同信噪比下的Lorenz混沌时间序列以及实际采集的具有混沌特性的温度时间序列进行单步预测,仿真表明其能够保证算法的稳定性和收敛性,不存在LMS算法和NLMS算法的发散问题.
In order to overcome the problems caused by improper parameter selection when estimating the second-order Volterra filter coefficients using Least Mean Square (LMS), Normalized LMS (NLMS) or Recursive Least Square (RLS) algorithm, a new algorithm based on posterior error assumption and variable convergence Second order Volterra adaptive filter (DFPSOVF) for the Davidon-Fletcher-Powell (DFP) method is proposed.The recursive updating formula of the autocorrelation inverse matrix estimation in the parameter estimation algorithm is given and the computational complexity of the algorithm is analyzed. The DFPSOVF filter is used to predict the Lorenz chaotic time series and the actual time series with chaotic characteristics. The simulation shows that the Lorenz chaos time series can ensure the stability and convergence of the algorithm, and there is no LMS algorithm And NLMS algorithm divergence problem.