基于FOCUSS的自适应去噪声误差消除方法

在煤矿监测监控信息融合系统研究中发现,冲击干扰造成的误差不仅容易造成误报警现象,而且严重影响了数据的预测精度和对生产安全状态的判断。为了解决这个问题,提出了基于FOCUSS的自适应去噪声学习算法,该算法根据误差的稀疏特性,利用盲信号分离方法的优势对误差信号进行提取,从而消除噪声对监测数据的影响。仿真实验将其与小波消除数据误差法进行了对比,验证了基于FOCUSS的自适应去噪声学习算法的有效性。

Adaptive de-noising and error elimination learning based on FOCUSS

Errors due to impact disturbances in the monitoring information fusion system of a coal mine reduce the data forecasting precision which affects production safety and result in false reports when the system is operating normally.An adaptive de-noising error learning method based on FOCUSS was developed to provide more accurate monitoring information.A blind signal separation method based on the sparse characteristics of the error signals was developed to extract the error signals and eliminate the influence of noise on the monitoring data.Simulations using FOCUSS show that the de-noising method more effectively eliminates sparse errors than the wavelet elimination data error method.