摘要
针对目前去趋势项所用方法中存在需预先假设趋势项类型、计算复杂等问题,提出了基于平滑先验法(SPA)的被动声信号趋势项消除方法。该方法是一种改进的消除非平稳趋势项算法,通过分析不同规则化参数下SPA的截止频率,并结合目标声信号中趋势项频率范围,得出规则化参数取值以达到有效去除待处理信号中趋势项的目的,同时保留感兴趣的模式。对装甲目标声信号仿真分析结果表明,处理前后目标声信号时域波形及PSD图谱中趋势项去除效果明显,简单有效地分离了原始数据趋势项和周期项,可用来对雷场中目标声信号进行去趋势项预处理。
Present methods of trend items removing has to presuppose the type of trend items and have a large calculation amount. Aiming at these problems, a new way of smoothness priors approach （SPA） was proposed in this paper. This method is an improved algorithm in removing non-stationary signal trend items . By analy- zing the cutoff frequency of SPA with different regularization parameter and combining with the frequency range of signal trend items, reasonable value of regularization parameter was pointed out, in the meantime retain the interesting pattern. Simulation analysis of target signal collected from OTC experiment was conducted. The re- sults showed that the trend items in the samples were effectively eliminated view from the time domain waveform and PSI） spectrum before and after the treatment, moreover the trend items and periodic items in primary data was simply and effectively separated, the algorithm could be used to eliminate the trend items of the target a- coustic signal in minefield.
出处
《探测与控制学报》
CSCD
北大核心
2015年第4期34-36,41,共4页
Journal of Detection & Control
基金
总装预研基金项目(ZLY2008424)
