正交小波变换下单调迭代算法仿真研究
|
摘要
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.
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.
引文
[1] 陈炎龙,段红玉,孔素真,等.一种分布式数据库查询优化算法的研究[J].科技通报,2017,33(2):86-89.
[2] 张余,等. 基于相关标识符的频谱水印嵌入与提取方法[J]. 电波科学学报, 2016,31(1):25-31.
[3] 梁志国. 正弦波形参量对ADC有效位数评价的影响[J]. 计量学报, 2017,38(1): 20-25.
[4] 卢其威,等. 固定频率正弦波输出LLCC串并联谐振电路研究[J]. 电机与控制学报, 2017,21(1):98-107.
[5] 凌同华,等. 双正交小波基构造法及其在爆破振动信号分析中的应用[J]. 振动与冲击, 2018,37(11):273-280.
[6] 王森,等. 基于反对称双正交小波变换的多尺度归一化分割方法[J]. 计算机辅助设计与图形学学报, 2016,28(1):106-118.
[7] 何琳,等. 自适应加权全变分的低剂量CT统计迭代算法[J]. 计算机应用, 2016,36(10):2916-2921.
[8] 邹时禧,等. 基于时域分段处理的单频信号检测算法分析[J]. 微电子学与计算机, 2017,34(4):60-64.
[9] 高明信. 运动疲劳过程中脑电信号特征提取仿真[J]. 计算机仿真, 2017,34(5): 277-280.
[10] 庞宝君,等. 碎片云超高速撞击声发射信号特征分析[J]. 高压物理学报, 2014,28(6): 664-670.
[2] 张余,等. 基于相关标识符的频谱水印嵌入与提取方法[J]. 电波科学学报, 2016,31(1):25-31.
[3] 梁志国. 正弦波形参量对ADC有效位数评价的影响[J]. 计量学报, 2017,38(1): 20-25.
[4] 卢其威,等. 固定频率正弦波输出LLCC串并联谐振电路研究[J]. 电机与控制学报, 2017,21(1):98-107.
[5] 凌同华,等. 双正交小波基构造法及其在爆破振动信号分析中的应用[J]. 振动与冲击, 2018,37(11):273-280.
[6] 王森,等. 基于反对称双正交小波变换的多尺度归一化分割方法[J]. 计算机辅助设计与图形学学报, 2016,28(1):106-118.
[7] 何琳,等. 自适应加权全变分的低剂量CT统计迭代算法[J]. 计算机应用, 2016,36(10):2916-2921.
[8] 邹时禧,等. 基于时域分段处理的单频信号检测算法分析[J]. 微电子学与计算机, 2017,34(4):60-64.
[9] 高明信. 运动疲劳过程中脑电信号特征提取仿真[J]. 计算机仿真, 2017,34(5): 277-280.
[10] 庞宝君,等. 碎片云超高速撞击声发射信号特征分析[J]. 高压物理学报, 2014,28(6): 664-670.