﻿ 船用行程传感器电磁干扰信号抑制算法研究
 舰船科学技术  2017, Vol. 39 Issue (9): 155-158 PDF

Research on the suppression algorithm of electromagnetic interference signal of stroke sensor in maritime navigation
HUANG Ying-bang, MA Sheng-wei, WU Qia-er
South China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Guangzhou 510300, China
Abstract: The development of navigation industry is closely related to the development of national economy. At present, many navigational sensors are very susceptible to electromagnetic interference in the working environment, which is very unfavorable to the development of the navigation industry. Based on the characteristics of electromagnetic signal when the marine stroke sensor is working, noise reduction and de-noising experiments are carried out on the output signal of the stroke sensor with different noise suppression algorithms in order to minimize the noise interference of the stroke sensor in the navigating vessel, so that the naval vessel can work normally.
Key words: travel sensor     electromagnetic interference signal     suppression algorithm
0 引　言

1 小波降噪算法 1.1 小波降噪原理

 $W{T_x}\left( {a,\tau } \right) = \frac{1}{{\sqrt a }}\int_{ - \infty }^{ + \infty } {x\left( t \right){\Psi ^ * }\left( {\frac{{t - \tau }}{a}} \right)} {\rm d}t,a > 0\text{，}$ (1)

 $s\left( k \right) = f\left( k \right) + e\left( k \right);k = 0,1, \cdot \cdot \cdot ,n - 1\text{。}$ (2)

 $S = a{L_N} + a\sum\limits_{i = 1}^N {{H_i}} \text{。}$ (3)
 图 1 噪声信号近似部分的N层分解 Fig. 1 The N layer decomposition of noise signal approximation

 $\left\{ {\begin{array}{*{20}{c}}{\mathop {{{\hat \omega }_{j,i}} = }\limits_{sign\left( {{w_{j,i}}} \right)\; \cdot }}\\0,\end{array}} \right.\begin{array}{*{20}{c}}{ \left( {\left| {{w_{j,i}}} \right|} \right),} & {\left| {{w_{j,i}}} \right| \geqslant \lambda }\\{} & {\left| {{w_{j,i}}} \right| < \lambda }\end{array},0 \leqslant \alpha \leqslant 1\text{。}$ (4)

1.2 行程传感器的输出信号

 图 2 传感器输出信号和量程的关系 Fig. 2 Relationship between output signal and range of sensor

 图 3 加噪后输出信号和量程的关系 Fig. 3 The relationship between output signal and range after adding noise
1.3 小波降噪原理在行程传感器中的应用

1）选取合适的小波函数

 $\Psi \left( k \right) = \sum\limits_k {{g_k}\varphi \left( {2t - k} \right)} \text{。}$ (5)

2）信号降噪仿真图

 图 4 小波分解后的信号降噪图 Fig. 4 Signal noise reduction after wavelet decomposition

 图 5 小波函数降噪处理所得信息曲线和误差曲线 Fig. 5 The information curve and error curve of wavelet denoising

1.4 仿真验证

 $e = \frac{1}{n}\sum\limits_{i = 1}^n {\sqrt {{{\left( {{{y'}_i} - {y_i}} \right)}^2}} } \text{，}$ (6)

n取值2 000时，计算可得误差均值是0.3%。因此借助软硬阈值折中法对输出信号实施小波降噪处理后，船用行程传感器的电磁噪声干扰信号得到明显抑制。

2 相位差计算算法

 $W\left( j \right) = 0.42 - 0.5\cos \left( {\frac{\pi }{{100}}j} \right) + 0.08\cos \left( {\frac{{2\pi }}{{100}}j} \right)\text{，}$ (7)

1）S1实部：

 ${S_{1Rc}} = \sum\limits_{i = 1}^{{K_2}} {\sum\limits_{j = 1}^{{K_1}} {{S_1}\left( {i + j} \right)} } {}^ \circ \cos \left( {\frac{{2\pi }}{{{T_1}}}{}^ \circ i} \right){}^ \circ W\left( j \right)\text{，}$ (8)

S1虚部：

 ${S_{1{\mathop{\rm Im}\nolimits} }} = \sum\limits_{i = 1}^{{K_2}} {\sum\limits_{j = 1}^{{K_1}} {{S_1}\left( {i + j} \right)} } {}^ \circ {\rm sim}\left( {\frac{{2\pi }}{{{T_1}}}{}^ \circ i} \right){}^ \circ W\left( j \right)\text{。}$ (9)

 $\tan {\theta _1} = \frac{{{S_{1{\mathop{\rm Im}\nolimits} }}}}{{{S_{1Rc}}}}\text{。}$

2）S2实部：

 ${S_{2Rc}} = \sum\limits_{i = 1}^{{K_2}} {\sum\limits_{j = 1}^{{K_1}} {{S_2}\left( {i + j} \right)} } {}^ \circ \cos \left( {\frac{{2\pi }}{{{T_1}}}{}^ \circ i} \right){}^ \circ W\left( j \right)\text{，}$ (10)

S2虚部：

 ${S_{2{\mathop{\rm Im}\nolimits} }} = \sum\limits_{i = 1}^{{K_2}} {\sum\limits_{j = 1}^{{K_1}} {{S_2}\left( {i + j} \right)} } {}^ \circ sim\left( {\frac{{2\pi }}{{{T_1}}}{}^ \circ i} \right){}^ \circ W\left( j \right)\text{，}$ (11)

 $\tan {\theta _2} = \frac{{{S_{2{\mathop{\rm Im}\nolimits} }}}}{{{S_{2Rc}}}}\text{，}$

 $\left| {{\theta _1} - {\theta _2}} \right| = \left| {\arctan \left( {\frac{{{S_{1{\mathop{\rm Im}\nolimits} }}{S_{2Rc}} - {S_{1Rc}}{S_{2{\mathop{\rm Im}\nolimits} }}}}{{{S_{1{\mathop{\rm Im}\nolimits} }}{S_{2Rc}} - {S_{1Rc}}{S_{2{\mathop{\rm Im}\nolimits} }}}}} \right)} \right|\text{。}$ (12)

3 综合算法

1）小波软硬阈值降噪算法处理信号时，当分级层数很低时，其信号降噪效果非常不好，当分解层数太高时会发生有用信号失真问题；

2）相位差算法处理信号时，会出现降噪效果不好的问题。因此本文基于2种算法原理，充分结合各种算法的特点，设计形成了一种综合算法，即对 2 组线圈采集获得的输入信号先利用小波软硬阈值降噪算法进行降噪处理，后再通过相位差计算算法对输出信号进行降噪处理。由于综合算法相比较2种单一算法在信号降噪处理时，过程更复杂，若参数还和2种单一算法处理时相同，将造成严重的信号失真。因此，适当调整综合算法参数：1）小波降噪分解层数为5层；2）相位差算法的相位宽度范围180°±30°。随机从采集数据中选取500组，采用上述包含自适应算法在内的4种降噪算法对输出信号进行降噪处理，对比结果详见表1。由结果可知，通过综合算法降噪处理的原始信号，其数据的良好率显著上升，且有用信号失真率明显降低，得到有效控制。

4 结　语

 [1] 张伟, 师奕兵, 卢涛. 小波神经网络在无线随钻测量系统在泥浆信号检测中的应用研究[J]. 电子测量与仪器学报, 2008 (6): 98–101. [2] 乐波, 刘忠, 古天祥. 一种低信噪比线性调频脉冲信号参数提取方法[J]. 电子测量与仪器学报, 2008 (5): 76–79. [3] 胡柏青, 魏峥, 王伯雄. 强噪条件下基于小波降噪的陀螺仪声信号处理方法[J]. 传感技术学报, 2008 (6): 81–84.