摘要
基于迭代最小二乘支持向量拟合(IRWLS-SVR)算法,建立了舰船噪声中线谱信号的数学模型用于对噪声背景中线谱信号的检测;利用IRWLS-SVR算法得到的稀疏解,减少了运算量,提高了运算速度。利用计算机仿真信号和实测舰船声信号的验证表明,该算法在高斯噪声背景下可以有效地检测出线谱信号,检测舰船噪声信号中线谱信号的信噪比改善达到4dB。
A mathematical model for the line spectrum signals of ship-radiated noise based on IRWLS-SVR algorithm is proposed, which is of high calculation speed and low computational cost because of the sparse solution. The simulation results by using measured noise data show that the algorithm can detect the line spectrum signal in the Gaussian noise background effectively, with about 4 dB signal to noise ratio improvement.
出处
《电声技术》
2010年第4期47-49,共3页
Audio Engineering
关键词
支持向量拟合
迭代权值最小二乘法
线谱
信号检测
support vector regression
Iterative re-weighted least squares
line spectrum signals
signal detection
