正交小波变换下单调迭代算法仿真研究

Simulation of Monotone Iterative Algorithm Based on Orthogonal Wavelet Transform

传统低阶单调迭代算法的扰动信号输出量拟合精度不足,算法收敛速度慢。为了解决上述问题,在正交小波变换条件下提出基于高阶的单调迭代算法研究。利用正交小波变换消噪,提取扰动信号单调迭代特征,对特征向量进行盲均衡处理。确定扰动信号单调区间内的梯度值和平滑系数,构造上下解向量,给出高阶条件下的联立方程组,并经过多次反复迭代求出上解和下解,逼近最优值,提高单调迭代算法的收敛性能。仿真结果表明,正交小波变换条件下提出迭代算法的扰动信号输出量拟合度可达到98.71%,迭代算法的收敛速度更快、CPU耗时更短。

This article puts forward a monotone iterative algorithm based on high order under the orthogonal wavelet transform. At first, the orthogonal wavelet transform was used to eliminate noise, and then the monotone iterative feature of disturbance signal was extracted. Moreover, the blind equalization was performed on feature vector. Meanwhile, the gradient value and the smoothness coefficient in the monotone interval of disturbance signal were determined, and then upper solution vector and lower solution vector were constructed. In addition, the simultaneous equations system under the high-order condition was given. After repetitious iterations, the upper solution and the lower solution were obtained, which converged to optimal value. Finally, the performance of monotone iterative algorithm was improved. Simulation results show that the fitting degree of quantity of disturbing signal output of proposed iterative algorithm can reach 98.71% in the orthogonal wavelet condition. The convergence speed of iterative algorithm is faster and the time consumption of CPU is shorter.