Machine Learning for Signal Demodulation in Underwater Wireless Optical Communications

The underwater wireless optical communication(UWOC)system has gradually become essential to underwater wireless communication technology.Unlike other existing works on UWOC systems,this paper evaluates the proposed machine learningbased signal demodulation methods through the selfbuilt experimental platform.Based on such a platform,we first construct a real signal dataset with ten modulation methods.Then,we propose a deep belief network(DBN)-based demodulator for feature extraction and multi-class feature classification.We also design an adaptive boosting(Ada Boost)demodulator as an alternative scheme without feature filtering for multiple modulated signals.Finally,it is demonstrated by extensive experimental results that the Ada Boost demodulator significantly outperforms the other algorithms.It also reveals that the demodulator accuracy decreases as the modulation order increases for a fixed received optical power.A higher-order modulation may achieve a higher effective transmission rate when the signal-to-noise ratio(SNR)is higher.

A new method of demodulating the digital frequency modulation signals

A new method of estimating the frequency-known signals from the strong background noise was presented first. Then the new method was used in the demodulation of the digital frequency modulation (FSK) signals. The new demodulation method can complete the demodulation of the FSK signals only with the carrier frequency and without any carrier phase information. The simulation results show that the performance of anti-noise of the new method is better than that of the incoherent demodulation method and the fluctuation of the carrier phase has little effect on the new method. So the new demodulation method has a fine prospect in the practical applications.

Performance Comparison of Three Algorithms Applied to UM2000 Signal Demodulation

UM2000 signal is a type of multi-audio frequency-modulated signal which is widely used for railway blocking. Principles of three typical demodulating algorithms are presented in details in this paper. Bit error rates of the three methods at different SNRs are achieved by Monte Carlo simulation experiments. Among the three algorithms, the quadrature demodulation has the best performance at the real working environment. However, the three methods have the same problem of phase hopping when noise is too strong.

Parameterized time-frequency analysis to separate multi-radar signals

Multi-radar signal separation is a critical process in modern reconnaissance systems. However, the complicated battlefield is typically confronted with increasing electronic equipment and complex radar waveforms. The intercepted signal is difficult to separate with conventional parameters because of severe overlapping in both time and frequency domains. On the contrary, time-frequency analysis maps the 1D signal into a 2D time-frequency plane, which provides a better insight into the signal than traditional methods. Particularly, the parameterized time-frequency analysis (PTFA) shows great potential in processing such non stationary signals. Five procedures for the PTFA are proposed to separate the overlapped multi-radar signal, including initiation, instantaneous frequency estimation with PTFA, signal demodulation, signal separation with adaptive filter and signal recovery. The proposed method is verified with both simulated and real signals, which shows good performance in the application on multi-radar signal separation.

A novel demodulation method for transmission using nitrogen–vacancy-based solid-state quantum sensor

Diamond based quantum sensing is a fast-emerging field with both scientific and technological significance.The nitrogen–vacancy(NV)center,a crystal defect in diamond,has become a unique object for microwave sensing applications due to its excellent stability,long spin coherence time,and optical properties at ambient condition.In this work,we use diamond NV center as atomic receiver to demodulate on–off keying(OOK)signal transmitted in broad frequency range(2 GHz–14 GHz in a portable benchtop setup).We proposed a unique algorithm of voltage discrimination and demonstrated audio signal transceiving with fidelity above 99%.This diamond receiver is attached to the end of a tapered fiber,having all optic nature,which will find important applications in data transmission tasks under extreme conditions such as strong electromagnetic interference,high temperatures,and high corrosion.

基于Labview的分布式光纤光栅信号解调技术研究

分布式光纤光栅信号解调中受到通信信道的码间干扰导致信道均衡性不好,提出基于Labview的分布式光纤光栅信号解调技术。采用连续时间采样技术构建分布式光纤光栅信号采集模型,对信号输出的功率谱密度特征进行相位调制处理,在离散的维纳模型下提取分布式光纤光栅信号的偏移中心频率和单边带功率谱密度特征信息,通过Labview的调制解调技术实现对信号的调制解调处理,最后采用抗相位噪声干扰的星座图分析方法分析光纤光栅信号解调的抗干扰性和误码抑制能力。仿真结果得知,该方法进行分布式光纤光栅信号解调的抗干扰性较好,提高了信道的频谱带宽,降低了频率中心偏移,传输误码率低于0.05%,信号的传输延时维持在30μs左右,具有较强的应用性。

高速精确的光纤光栅信号解调方案设计

为提升光纤光栅的应用效果,为此,设计高速精确的光纤光栅信号解调方案。依据LPFG的光纤光栅信号解调原理,设计两个LPFG双边缘滤波器处理两路反射光,并且利用光电探测器探测光功率结果,以此解调光栅的中心波长;采用光纤干涉仪检测解调频率,对其进行正反扫描,依据光传播时延引起的解调误差和解调频率扫描方向之间的关联,完成解调结果补偿。测试结果显示:该方案具有较好可行性,光谱发生不同程度重叠时,光纤光栅信号的解调结果标准差均小于1.85 pm;光栅波长的解调误差均在-5~5 pm之间;有效降低解调误差,提升光纤光栅解调精度,可靠获取信号中的信息。

光学多普勒差分流速仪数据解调方法验证

光学多普勒差分流速仪在测量海水流速时,由于信号微弱和噪声干扰,接收信号信噪比很低,信号解调面临挑战,可能导致较大的测量误差。为了准确解调埋藏在噪声中的信号并获取速度信息,需要对获取的海试测量数据进行进一步算法处理。采用自适应滤波算法有效降低了多普勒信号中混杂的高频噪声干扰,提高了海试原始数据频域信号的信噪比,并利用寻峰算法更准确地获取了谱峰值。结果表明,与直接读取原始数据的谱峰值相比,经过该算法处理后,激光差分多普勒流速仪与声学多普勒流速仪比测的流速拟合度平均误差为0.2163 cm/s,误差降低了25.5%,表明该算法对激光多普勒测速信号解调是有帮助的。

基于振动信号的控制棒驱动机构滚轮早期故障诊断研究

针对现有研究未充分关注控制棒驱动机构(control rod drive mechanism,CRDM)的早期故障诊断问题、很难将故障特征定位至具体部件以及人工引入的故障样本与装备实际故障特征存在差异等不足,提出了一种基于振动信号的CRDM滚轮早期故障诊断方法:首先,利用寿命考核试验时机采集了某密封磁阻马达式CRDM的滚轮全寿命振动信号,基于经验模态分解(empirical mode decomposition,EMD)和Hilbert变换方法进行解调分析,获得与滚轮退化状态相关的模态成分;然后,采用时、频域分析方法获得了11个能够直接表征CRDM滚轮磨损状态的特征量,并根据退化趋势提取出与实际故障特征高度吻合的早期故障样本;最后,分别基于BP神经网络和支持向量机两种方法实现了CRDM滚轮早期故障的多特征智能诊断。结果表明:提取的滚轮早期磨损故障样本与实际运行过程保持了较好的一致性,证明所提CRDM滚轮早期故障诊断方法具有较强的工程应用价值。

一种差动式位移传感器芯片的信号调理电路设计

设计了一种适用于线性可变差动式位移传感器(LVDT)的信号调理电路,该电路用于将LVDT的副边输出信号进行处理,并输出一个表征铁芯位移的直流电压;主要由解调电路、除法电路和运算放大器组成。LVDT副边输出为两个互补的正弦波信号,直接驱动信号调理电路;由于该信号调理电路采用了独特的比值架构,得到的是两个电压的差与和的比值,所以不要求LVDT初级激励使电压幅度保持稳定,同时,LVDT初级与次级线圈的相移以及干扰电压对该电路的功能不会产生影响,能有效地解决传统信号调理电路在测量精度和可靠性等方面的问题,具有更好的稳定性和性能优势,经过仿真验证及流片测试表明,该信号调理电路总误差≤1.49%/FS,线性度≤0.05%,满量程输出电压≥±10 V。各项功能指标均达到设计要求。

基于解调重构的跳频信号辐射源个体识别方法

现有跳频信号辐射源个体识别方法大多围绕跳变瞬态的特征进行讨论,其捕获和精准定位的难度较大。因此,提出了一种基于解调重构的跳频信号辐射源个体识别方法,能够有效利用跳频信号的稳态信息进行辐射源个体识别。首先对跳频信号进行跳变定时,提取出各跳的基带波形;然后解调出各跳的符号,并将其经过理想成型滤波器,得到理想的重构基带波形;最后将原始波形和重构波形一起送入神经网络,得到分类结果。给出发射机畸变模型和畸变参数范围,经过多次测试给出网络参数的建议取值范围,并进行验证实验。实验结果表明,该方法能够有效地完成跳频信号辐射源个体识别任务。

基于磁编码器和窄带解调的太阳轮故障检测

绝对式磁编码器具有价格低、安装方便等优势,适合于工程应用,但在低转速工况下其角度信号受测量误差的影响较大,需要对角度信号进行拟合以消除测量误差带来的干扰。传统的分段拟合方法在分段点处易出现不连续现象,角度信号的不连续导致由其计算的瞬时角速度(instantaneous angular speed,简称IAS)信号中出现虚假冲击,并对基于IAS信号的故障特征提取造成干扰。针对此问题,提出一种基于绝对式磁编码器IAS估计和窄带解调的太阳轮故障检测方法。首先,为保证IAS信号的连续性与平滑性,提出基于最小二乘拟合与5次Hermite插值的方法,实现了对IAS信号的拟合估计和平滑处理;其次,结合窄带解调技术实现了对行星齿轮减速器太阳轮故障特征的有效提取;最后,通过对正常与故障太阳轮的绝对式磁编码器信号分析对比,验证了所提方法的有效性。

基于IAS信号自适应窄带解调的RV减速器摆线轮故障特征提取

RV(rotary vector)减速器广泛应用于工业机器人,研究其关键零部件的故障检测方法具有重要意义。由于RV减速器结构复杂、摆线轮转速低、故障冲击较弱等问题,造成摆线轮故障特征难以用常规振动分析方法提取。针对上述问题,提出一种基于瞬时角速度信号自适应窄带解调的RV减速器摆线轮故障特征提取方法。首先对编码器角度信号进行采集,然后使用向前差分法获得瞬时角速度信号,在此基础上,使用角域同步平均提取摆线轮周期相关同步分量,消除非同步干扰,并通过构建边带信噪比指标选取故障信息丰富的啮合谐波频带进行带通滤波,最后对滤波后的信号进行窄带解调,实现对RV减速器摆线轮的故障特征提取。试验结果表明,该方法可以有效地提取RV减速器摆线轮故障特征,且自适应实现优化带宽选择。

An improved initial rotor position estimation method using high-frequency pulsating voltage injection for PMSM

High frequency pulsating voltage injection method is a good candidate for detecting the initial rotor position of permanent magnet synchronous motor.However,traditional methods require a large number of filters,which leads to the deterioration of system stability and dynamic performance.In order to solve these problems,a new signal demodulation method is proposed in this paper.The proposed new method can directly obtain the amplitude of high-frequency current,thus eliminating the use of filters,improving system stability and dynamic performance and saving the work of adjusting filter parameters.In addition,a new magnetic polarity detection method is proposed,which is robust to current measurement noise.Finally,experiments verify the effectiveness of the method.

基于自适应极值的白光干涉信号解调算法研究

针对传统时域解调算法等光程点定位稳定性差、抗干扰能力弱,而频域解调算法效率低且部分方法受扫描采样步长限制等问题,课题组提出一种基于自适应极值的白光干涉解调算法。从干涉信号统计学规律出发,通过对干涉信号包络曲线加权平均,将它修正为高斯型包络曲线;然后,通过对包络曲线高斯拟合插值得到等光程点,进而得到测量方向的深度信息以及三维形貌;最后,以高度为20μm标准台阶样块为测量对象进行多次实验。结果表明:与传统算法相比,课题组所提出算法的测量重复性和精度均有显著提升。该算法为白光干涉中信号解调提供了一种重复性更好、误差更小的方法。

基于数据驱动的柱塞泵冲击信号分离模型研究

针对多缸体柱塞泵冲击信号难以分离且机械故障判断困难的问题,提出了基于数据驱动的柱塞泵冲击信号分离模型。模型根据谱峭度最大原则对原始信号进行自适应带通滤波处理,根据二次包络谱峰值结合最优谐波能量和算法自动搜索柱塞泵的转速数据,根据转速对柱塞泵的单次循环大周期进行分割,以提取每个大周期内的各个密封阀单次冲击信号;判断每次单个冲击的起始时刻,设计周期时变滤波器组,对原始信号进行滤波处理,得到单一缸体的冲击信号,并计算相应的特征指标;采用实际柱塞泵振动数据对模型的准确性进行了验证。结果表明:对于具有明显冲击的三缸泵振动信号,采取分离模型分析出其转频为3.7566 Hz,信号周期为5.12 s,设计出6个周期时变滤波器组对原始信号进行处理,实现了每个密封阀的冲击信号的有效分离,便于后续对每个阀体进行独立监测和诊断分析。研究结果可为多源冲击信号的分离提取提供技术支持以及参考依据。

光栅型缸压传感器解调系统设计

为使携带被测量信息的光信号更加真实地还原缸压信号,从光纤感知原理出发分析解调系统构成和工作过程,设计解调缸压信号的电路结构并分析研究参数指标取值。利用Cadence PSpice仿真软件及电路图制版实物试验验证了所做工作的合理性。对所设计电路进行灵敏度分析,确定对电路工作性能影响较大的关键元器件,并预测关键器件取不同容差±5%、±1%、±0.5%、±0.1%时电路性能指标的变化范围,可为研究人员选择最佳器件提供依据。

基于DPCA的涡旋压缩机主轴故障信号调制特性研究

为了揭示涡旋压缩机的振动和噪声的机制与调制特性,通过搭建实验平台,模拟主轴磨损故障。利用信号解调方法对振动及噪声信号特性进行讨论,探究此类故障对涡旋压缩机的影响。通过与循环平稳分析方法对比,验证了基于时频分析与主成分分析(DPCA)的信号解调方法处理涡旋压缩机信号的优越性。利用DPCA方法提取涡旋压缩机的信号调制特征,讨论了不同工况下的信号特性及产生机制。结果表明:周期性变化的电磁力与压力增强了主成分中调制频率的幅值,x和y方向的振动对噪声的变化起主导作用。

基于BOTDR分布式检测技术的光缆隐蔽性缺陷识别

提出基于布里渊光时域反射(Brillouin optic time domain reflectometer,BOTDR)分布式检测技术与解调信号的光缆隐蔽性缺陷识别方法,识别检测光缆隐蔽性缺陷。在生产阶段将光纤植入多股碳纤维导线复合芯内部作为传感器,依据BOTDR技术与光时域反射(optic time domain reflectometer,OTDR)技术原理,构建关于碳纤维导线光缆隐蔽性检测的分布式传感系统,利用光纤对温度、应力、传播损耗的高精度感知,多维度地分析碳纤维导线光缆的隐蔽性缺陷及位置分布,实现光缆的隐蔽性缺陷识别。利用morlet小波解调布里渊散射信号,提取包络信息,去除信号噪声,采用列文伯格-马夸尔特算法,拟合布里渊散射谱数据,精确估计最优Brillouin频移量参数,提升整体缺陷识别精度。实验证明:该方法可以准确地对光缆的隐蔽性缺陷进行表征,实现多股碳纤维导线光缆隐蔽性缺陷的快速、有效识别,促进碳纤维导线光缆更好地应用于增容、大跨越等工程中。

光纤传感器非周期低频动态信号相位载波解调方法研究

为了提高光纤传感器非周期低频动态信号的相位载波解调效果,提出光纤传感器非周期低频动态信号相位载波解调方法研究。采用自适应波束形成算法增强非周期低频动态信号的细节信息;通过奇异值分解算法提取信号的特征;将提取的特征输入到基于双扩展卡尔曼滤波的CD3S信号解调模型中,通过卡尔曼滤波结构剔除噪声;通过联合估计实现动态信号的扩频码同步,完成光纤传感器非周期低频动态信号相位载波解调。实验结果表明,所提方法在有无噪声干扰下解调误码率均处于1.0%以下,解调后位移误差低至12 pm,提高了解调效果和抗噪声能力。