司法消噪算法的非线性改进研究

Improvement of Judicature De-Noising Algorithm Through Nonlinear Optimization

研究智能人工呼吸床信号优化设计问题,针对二期智能人工呼吸床系统信号预测难,采用司法消噪算法的法规综合函数带来的线性弊端,进行了非线性优化研究。为了准确预测在临床实验的原算法消噪效果分析基础上,结合人工神经网络良好的非线性预测功能,建立了关于RBF神经网络的法规综合指数模型,并与原始算法的线性法规综合函数进行动态结合,不仅有效提高了噪声评判标准的准确度,同时也保证了实际应用的可行性。利用matlab平台对实验数据进行仿真,结果表明,改进算法在保证良好去噪效果的前提下,效果明显优于原算法,且能快速反应陡升陡降的新稳态变化。

Aiming to solve the problems of the linear function of the Judicature De-Noising algorithm in the second phase of intelligent artificial respiration bed system,we optimized it by nonlinear function in this paper.Combined with the neural network,we set up a RBF regulations composite index model on the basis of the original algorithm＇s clinical test.And link to the linear regulations composite index of original algorithm,it can not only improve the accuracy of noise evaluation standards efficiently,but also ensure the feasibility of practical application.Simulating with matrix laboratory software,the results show that the algorithm can accurately distinguish the signal and noise,and response the change of new steady state rapidly.