基于非线性样条滤波算法抑制网络噪声干扰

Network Noise Interference is Suppressed Based on Nonlinear Spline Filtering Algorithm

在网络传输的过程中,传输的信号数据会受到各种外来或者自身数据损耗引发安全问题。为解决信号数据受到的外部干扰,基于滤波算法的研究受到了广泛关注。传统的系统,一般是基于有限脉冲响应滤波器构建的线性系统,但在实际环境中,系统一般是以非线性的形式存在的。为更有效地解决系统中的网络安全问题,文章提出了一种新的非线性滤波算法,该非线性滤波算法由滤波过程和自适应过程组成。滤波算法中的滤波过程由传统的线性FIR滤波算法结合非线性模块构成,传统的FIR滤波算法迭代更新滤波系统的权重,并且使用随机梯度算法来自适应地改变非线性样条控制点。滤波过程的非线性部分通过Catmull-Rom样条来对控制点参数施加简单的约束。随后依据设计的自适应算法,探索了计算复杂度和滤波算法收敛时步长边界的选择范围,进一步验证试验,结果证明是有效的。仿真结果表明了所提出方法的有效性,该算法能够有效缓解网络安全问题,获得了可观的滤波效果。

In the process of network transmission,the transmitted signal data will be subject to various security issues caused by external or self-data loss.In order to solve the external interference of signal data,the research based on filtering algorithm has received extensive attention.Traditional systems are generally linear systems based on finite impulse response filters.But in the actual environment,the system generally exists in a nonlinear form.In order to solve the network security problem in the system more effectively,in this paper,we propose a new nonlinear filtering algorithm,which consists of a filtering process and an adaptive process.The filtering process in the filtering algorithm is composed of a traditional linear FIR filtering algorithm combined with a nonlinear module.The traditional FIR filtering algorithm iteratively updates the weights of the filtering system,and wherein a stochastic gradient algorithm is used to adaptively change the nonlinear spline control points.The nonlinear part of the filtering process uses Catmull-Rom splines to impose simple constraints on the control point parameters.Then,based on the adaptive algorithm we designed,we explored its computational complexity and the selection range of the step size boundary conditions when the filtering algorithm converges.Further,experiments are carried out to verify that the results are effective.The simulation results show the effectiveness of the proposed method.The algorithm can effectively alleviate the network security problem and obtain considerable filtering effect.