【摘要】 瞬态干扰脉冲信号可使电子系统功能混乱,计算机死机,通讯、导航等系统紊乱甚至烧毁电子元器件等。通过采取屏蔽、滤波、旁路和瞬变电压抑制等硬件加固措施,可以使干扰信号不至于对仪器设备造成“硬损伤”,但信号可能已经失真,需要采取一定的措施来恢复信号。应用混合专家网络模型,选用能够提高网络学习效率的优化目标函数,可以有效地解决传统信号处理方法难以解决的参杂信号恢复问题,通过仿真试验也证明了ME模型改进算法与常用的FFT低通滤波方法处理特定掺杂信号的效果差异。更多还原

【Abstract】 Transient interfering electromagnetic impulse signal can lead to electronic system’s function confusion, computer corruption, communicating, navigation system mess or even destroying electronic components etc. By adopting shield, filtration, bypass , transient impulse suppressing and other hardware reinforcement measure, we can make the interference signal not result in harware damage, but the signal may have already distorted, and have to a-dopt certain measure to recover the signal. By applying mixture of experts net model, selecting optimization target functon that may improve net’s learning efficiency,we can effectively solve the mixed signal’s recovery that traditional signal’s procesing is difficult to cope with. According to emulation example, we can find out the different effects on handling the particular highly mixed signal by using the ME model’s improved arithmetic and the commonly used FFT lowpass filter method.更多还原