基于小波核SVM的水声信道盲均衡算法

Underwater Acoustic Channel Blind Equalizer Based on Wavelet Kernel SVM

推导并提出了一种快速收敛的基于小波核支持向量机的恒模盲均衡器(WSVM-CMBE)。该方法以支持向量机为框架,利用信号的恒模特性构造代价函数,采用小波核函数,并自适应调整核函数中的伸缩因子。通过水声信道仿真实验,并与采用高斯核函数的支持向量机恒模盲均衡器(SVM-CMBE)进行比较,结果证明该方法提高了收敛速度。

This paper presents a wavelet kernel function support vector machine-based constant modulus blind equalizer（WSVM-CMBE） with convergence speediness capability.The method exploits signal constant modulus feature to constructs the cost function,adopted wavelet kernel function of which the flex parameter is self-adaptively adjusted.Simulation is carried out to compare it with support vector machine-based constant modulus blind equalizer with Gauss kernel function（SVM-CMBE） .The result shows the method has much better performance on convergence speed.