WO2016082562A1 - 基于无线电信号频谱特征模板的信号识别方法及系统 - Google Patents

基于无线电信号频谱特征模板的信号识别方法及系统 Download PDF

Info

Publication number
WO2016082562A1
WO2016082562A1 PCT/CN2015/083989 CN2015083989W WO2016082562A1 WO 2016082562 A1 WO2016082562 A1 WO 2016082562A1 CN 2015083989 W CN2015083989 W CN 2015083989W WO 2016082562 A1 WO2016082562 A1 WO 2016082562A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
template
identified
bandwidth
difference
Prior art date
Application number
PCT/CN2015/083989
Other languages
English (en)
French (fr)
Inventor
张奇勋
冯志勇
王宝聪
高超
Original Assignee
北京邮电大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京邮电大学 filed Critical 北京邮电大学
Publication of WO2016082562A1 publication Critical patent/WO2016082562A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing

Definitions

  • the present invention relates to the field of wireless communication technologies, and in particular, to a signal recognition method and system based on a radio signal spectrum feature template.
  • Radio signal identification technology is widely used in the field of interference source investigation, which can make interference detection faster.
  • the main research direction of radio identification technology is to extract signal characteristics (center frequency, bandwidth, code rate, modulation mode, etc.) and use pattern recognition to identify signals.
  • the signal recognition objects in the existing research are all modulated signals, and do not take into account the impact of the actual more complex wireless communication environment on the signal.
  • the patent "a method of signal identification” constructs a signal template by using the morphological envelope of various signal power spectra. The complex representation of the template makes the signal recognition process cumbersome and difficult to apply in the implementation process.
  • the existing signal identification method is to extract the signal features, and to eliminate the uncorrelated signals step by step according to different signal characteristics, and finally obtain the signal recognition result, and the recognition process is cumbersome.
  • the existing signal identification is for modulated signals, such as Quadrature Phase Shift Keying (QPSK) signals, Quadrature Amplitude Modulation (QAM) signals, etc., without considering the actual signal. The effect of differences in encoding, transmission information, etc. on the signal spectrum.
  • QPSK Quadrature Phase Shift Keying
  • QAM Quadrature Amplitude Modulation
  • the present invention provides a signal recognition method and system based on a radio signal spectrum feature template, which can represent a plurality of types of signals through a spectrum template, and further simplify the signal recognition algorithm on the basis of ensuring the accuracy of signal recognition. .
  • the present invention provides a signal recognition method based on a radio signal spectral feature template, the method comprising: constructing a signal template library by extracting spectral features of a plurality of radio signals; and according to the signal template in the signal template library
  • the representation form is to preprocess the identification signal; compare and match the pre-processed signal to be identified with the signal template of the signal template library to obtain the type and characteristic parameter of the to-be-identified signal.
  • the signal template library is constructed by extracting spectral features of a plurality of radio signals, including: filtering, sampling, and transforming a plurality of radio signals to obtain a frequency domain signal; and performing frequency domain signals by using a domain averaging method. Smoothing processing; quantizing the smoothed signal to form a staircase signal to obtain a signal template.
  • the preprocessing the signal to be identified according to the representation form of the signal template in the signal template library comprises: filtering the signal to be identified according to a preset bandwidth; smoothing and quantizing the filtered signal , obtain the step signal after preprocessing.
  • the comparing and matching the pre-processed signal to be identified with the signal template of the signal template library to obtain the type and characteristic parameters of the to-be-identified signal includes: obtaining a center frequency of the to-be-identified signal And a difference between a center frequency of each signal template in the signal template library, if the difference is greater than a first preset value, determining that the to-be-identified signal does not match the signal template; if the difference is less than Determining, by the first preset value, a difference between a preset bandwidth of the to-be-identified signal and a preset bandwidth of the signal template, and if the difference is greater than a second preset value, determining the to-be-identified signal and the signal
  • the template does not match; if the difference is less than the second preset value, calculating a comprehensive deviation value of the to-be-identified signal and the signal template, and if the comprehensive deviation value is greater than a third preset value, determining the to-be-determined
  • the combined deviation value is:
  • r1 is the ratio of the difference between the power of the 3dB bandwidth of the signal to be identified and the signal template and the power value within the 3dB bandwidth of the signal template
  • r2 is within 10 dB of the signal to be identified and the signal template.
  • r3 is the ratio of the power mean square error of the signal to be identified and the signal template in the 10 dB bandwidth to the power value within the 10 dB bandwidth of the signal template.
  • the present invention provides a signal recognition system based on a radio signal spectral feature template
  • the system includes: a template construction unit, configured to construct a signal template library by extracting spectral features of the plurality of radio signals; and a preprocessing unit configured to preprocess the identification signal according to the representation form of the signal template in the signal template library; And a matching unit, configured to compare and match the pre-processed signal to be identified with the signal template of the signal template library to obtain a type and a feature parameter of the to-be-identified signal.
  • the template construction unit is specifically configured to: filter, sample, and transform a plurality of radio signals to obtain a frequency domain signal; perform smoothing on the frequency domain signal by using a domain averaging method; and smooth the processed signal A quantization process is performed to form a staircase signal to obtain a signal template.
  • the pre-processing unit is specifically configured to: filter the signal to be identified according to a preset bandwidth; perform smoothing and quantization processing on the filtered signal to obtain a pre-processed step signal.
  • the matching unit is configured to: obtain a difference between a center frequency of the to-be-identified signal and a center frequency of each signal template in the signal template library, and if the difference is greater than a first preset value, And determining that the to-be-identified signal does not match the signal template; if the difference is less than the first preset value, obtaining a difference between the preset bandwidth of the to-be-identified signal and the preset bandwidth of the signal template, if If the difference is greater than the second preset value, determining that the to-be-identified signal does not match the signal template; if the difference is less than the second preset value, calculating a combination of the to-be-identified signal and the signal template a deviation value, if the comprehensive deviation value is greater than a third preset value, determining that the to-be-identified signal does not match the signal template; and if the comprehensive deviation value is less than a third preset value, determining the to-be-identified A signal is matched with the
  • the combined deviation value is:
  • r1 is the ratio of the difference between the power of the 3dB bandwidth of the signal to be identified and the signal template and the power value within the 3dB bandwidth of the signal template
  • r2 is within 10 dB of the signal to be identified and the signal template.
  • r3 is the ratio of the power mean square error of the signal to be identified and the signal template in the 10 dB bandwidth to the power value within the 10 dB bandwidth of the signal template.
  • the present invention provides a signal recognition method and system based on a radio signal spectrum feature template, which is uniquely effective by extracting an image feature of a radio signal itself and constructing a spectrum template using a specific frequency and power point of the signal spectrum. Characterizing different kinds of radio signals can further simplify the signal recognition algorithm based on the accuracy of signal recognition.
  • FIG. 1 is a schematic flow chart of a signal identification method based on a radio signal spectrum feature template according to an embodiment of the present invention.
  • FIG. 2 is a schematic flowchart of a method for constructing a signal template library according to another embodiment of the present invention.
  • FIG. 3 is a schematic diagram of a signal spectrum template provided by another embodiment of the present invention.
  • FIG. 4 is a schematic diagram of a signal spectrum template step selection according to another embodiment of the present invention.
  • FIG. 5 is a schematic flow chart of a method for signal identification and matching according to another embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of a signal recognition system based on a radio signal spectrum feature template according to an embodiment of the present invention.
  • FIG. 1 illustrates a signal identification method based on a radio signal spectrum feature template according to an embodiment of the present invention.
  • the method includes the following steps:
  • Step 101 Construct a signal template library by extracting spectral features of various radio signals.
  • the signal template library does not need to be built every time, and a pre-built signal template library can also be used.
  • Step 102 Perform pre-processing on the identification signal according to the representation form of the signal template in the signal template library.
  • Step 103 Compare and match the pre-processed signal to be identified with the signal template of the signal template library to obtain a type and a feature parameter of the to-be-identified signal.
  • the process of constructing the signal template library in step 101 includes the following steps:
  • Step 201 Filter, sample, and transform a plurality of radio signals to obtain a frequency domain signal.
  • the radio signal is from the device transmitting signal, and after being filtered, sampled and transformed, the frequency domain signal is obtained.
  • the conversion must be the same as the number of points (such as 1024 FFT).
  • Step 202 Perform smoothing on the frequency domain signal by using a domain averaging method.
  • the jitter of the signal template in the frequency domain is very severe. If the signal template is directly drawn by this signal, the morphological similarity of the spectral template of the same signal is greatly reduced, thereby affecting the accuracy of signal recognition. To do this, the frequency domain signal needs to be smoothed.
  • the neighborhood averaging method in digital image processing is employed as the method of signal smoothing processing.
  • the smoothing filtering of the neighborhood averaging method adds the gray value of one pixel in the original image to the gray value of the neighboring pixels around it, thereby reducing the influence of noise on the image, as shown in the expression (1).
  • g(i,j) is the smoothed gray value
  • S is the set of neighbor points of the point to be processed
  • N is the number of points in S
  • f(i,j) is the image signal
  • n(i,j ) is a noise signal
  • F(i, j) is a noisy image signal.
  • the neighborhood usually takes 8 points around the processing point.
  • the amplitude curve of the frequency domain signal can be expressed as a function of frequency, so the neighborhood point takes two points around the frequency point.
  • the smoothing filtering method in the present invention includes the following two types, such as expressions (2) and (3). Where S'(n) is the smoothed signal amplitude; S(n) is the original signal amplitude.
  • the smoothing performance of the two is not much different.
  • the smoothing performance of the expression (3) is significantly better than the method of the expression (2). Therefore, the present invention selects the smoothing processing of the frequency domain signal.
  • the expression (4) is further derived for the expression (3). That is, for a spectrum signal with a unit of dBm, the average value can be smoothed.
  • the present invention proposes that the maximum jitter of the signal does not exceed 3 dB, that is, the signal is monotonous, and when there is jitter somewhere, the ratio of the amplitude between the peak of the jitter and the valley is not more than 3 dB.
  • Step 203 Perform quantization processing on the smoothed signal to form a staircase signal, and obtain a signal template.
  • the smoothed template signal is quantized into a staircase signal, which is a signal template.
  • the quantization used here is non-uniform quantization. Considering that most of the signals change slowly near the peak, the height of the step needs to be smaller. Otherwise, the sampling point corresponds to the upper frequency interval and the uniformity cannot be guaranteed.
  • the signal amplitude decreases, the signal waveform enters the steep segment.
  • the step height at this time needs to select a relatively large step; when the signal amplitude gradually becomes flat, the height of the step needs to be adjusted to be relatively small.
  • the steps in the signal template construction process can be selected as 1dB, 2dB, 3dB, 4.5dB, 6dB, 8dB, 10dB, 13dB, 16dB, 19dB, 22dB, 25dB, 28dB, 31dB, 34dB, 37dB, 40dB.
  • the signals of all template libraries are simplified to build the template for the sampling points according to the above quantitative criteria. As shown in Figure 4.
  • the quantized template signal is a stepped signal, and the signal template can be represented by a series of frequency sequence ⁇ f 1 , f 2 , f 3 ,..., f 35 ⁇ , considering that when the signal power is small, due to noise In the process of constructing a signal template, the signal can not always be reduced to -40 dB, and the values of some edge frequencies (such as f 1 , f 35 , etc.) are null. Therefore, the signal template is constructed as shown in Expression (5).
  • n is the number of FFT transform points
  • p 1 , p 2 , p 3 , ..., p 35 are power points corresponding to the respective frequency points
  • P 1 , P 2 , P 3 , ..., P n are power values.
  • the corresponding power points are also the same. This ensures the consistency of the spectrum template.
  • the preprocessing process of the signal to be identified in step 102 is as follows:
  • the signal to be identified is filtered according to a preset bandwidth; the filtered signal is smoothed and quantized to obtain a pre-processed step signal.
  • the signal preprocessing method proposed in this embodiment preprocesses the signal to be identified so that the signal to be identified has a similar representation to the signal template. Therefore, the signal preprocessing process is similar to the template construction process. However, since the signal is known during the template construction process, the filter bandwidth is also determined when filtering the signal. When the signal to be identified is filtered, since the signal type is unknown, the bandwidth for filtering the signal can be determined as the 40 dB bandwidth of the signal to ensure that the bandwidth of the signal to be identified is not lower than the signal bandwidth required by the spectrum template. The filtered smoothing and quantization is the same as the template construction process.
  • the method for identifying and matching the signal in step 103 includes the following steps:
  • Step 501 Obtain a difference between a center frequency of the to-be-identified signal and a center frequency of each signal template in the signal template library.
  • Step 502 Determine whether the difference obtained in step 501 is greater than the first preset value. If yes, go to step 508; otherwise, go to step 503.
  • Step 503 Obtain a difference between a preset bandwidth of the to-be-identified signal and a preset bandwidth of the signal template.
  • Step 504 Determine whether the difference obtained in step 503 is greater than a second preset value. If yes, go to step 508; otherwise, go to step 505.
  • Step 505 Calculate a comprehensive deviation value between the to-be-identified signal and the signal template.
  • Step 506 Determine whether the integrated deviation value is greater than the third preset value. If yes, go to step 508; otherwise, go to step 507.
  • Step 507 Determine that the to-be-identified signal matches the signal template, and obtain a type and a feature parameter of the to-be-identified signal.
  • Step 508 Determine that the to-be-identified signal does not match the signal template.
  • the difference between the center frequency of the signal to be identified (ie, f 18 ) and the center frequency of the signal template is obtained by calculation. If the center frequency of the signal to be identified deviates from the 10 dB bandwidth range of the signal template (ie, f 10 to f 26 ), the template is considered to be mismatched with the signal to be tested. Further, by comparing the difference between the 3dB bandwidth of the signal to be identified and the signal template (ie, f 22 -f 14 ), if the difference between the two exceeds 5%, the signal template is excluded from the effective signal set.
  • the comprehensive deviation value includes the following three cores. index:
  • the ratio of the difference between the power of the 3dB bandwidth of the signal to be identified and the signal template to the power value of the 3dB bandwidth of the signal template is set to r1; the representation of the signal template and the signal to be identified are In the case, the power within the 3dB bandwidth of the signal is as shown in the expression (6).
  • ⁇ 3 B 3 ⁇ 10 -0.3 + B 2 ⁇ (10 - 0.2 -10 -0.3 ) + B 1 ⁇ (10 - 0.1 -10 -0.2 ) + B 0 ⁇ (1-10 - 0.1 )
  • the power in the 3dB bandwidth of the signal to be identified and the signal template is determined by the expression (6), and the calculation method of the ratio r1 is as shown in the expression (7).
  • the ratio of the difference between the power of the 10dB bandwidth of the signal to be identified and the signal template to the power value of the 10dB bandwidth of the signal template is set to r2; similar to the calculation of the power within the 3dB bandwidth, the power calculation method of the 10dB bandwidth of the signal can be obtained.
  • the expression (8) is shown.
  • the power value of the 3dB bandwidth of the signal to be identified and the signal template can be determined by using the expression (8), and the calculation method of the ratio r1 is as shown in the expression (9).
  • the ratio of the power mean square error (matching filtering) of the to-be-identified signal and the signal template in the 10 dB bandwidth to the power value in the 10 dB bandwidth of the signal template is set to r3, as shown in the expression (10).
  • index 10l and index 10r are the left and right boundaries of the 10 dB bandwidth, respectively, that is, the numbers corresponding to p 10 and p 26 .
  • the comprehensive deviation index r abs(r1)+abs(r2)+abs(r3), and calculate the comprehensive deviation index of the signal to be identified and all template signals.
  • the integrated deviation index r of the signal to be identified is the smallest, the corresponding signal template is the most
  • a similar signal template further examines the degree to which the signal to be identified is similar to the most similar signal template.
  • the signal recognition method based on the radio signal spectral feature template provided by the embodiment can uniquely characterize different types of radio signals by extracting the image characteristics of the radio signal itself and constructing the spectrum template by using the specific frequency and power point of the signal spectrum.
  • the signal recognition algorithm is further simplified on the basis of ensuring the accuracy of signal recognition.
  • FIG. 6 shows a signal recognition system based on a radio signal spectrum feature template according to an embodiment of the present invention.
  • the system includes a template construction unit 601, a pre-processing unit 602, and a matching unit 603.
  • the template construction unit 601 is configured to construct a signal template library by extracting spectral features of the plurality of radio signals, and the pre-processing unit 602 is configured to perform pre-processing on the identification signal according to the representation form of the signal template in the signal template library.
  • the matching unit 603 is configured to compare and match the pre-processed signal to be identified with the signal template library to obtain a type and a feature parameter of the to-be-identified signal.
  • the template construction unit 601 is specifically configured to: filter, sample, and transform a plurality of radio signals to obtain a frequency domain signal; perform smoothing on the frequency domain signal by using a domain averaging method; and quantize the smoothed signal. Processing, forming a staircase signal, and obtaining a signal template.
  • the pre-processing unit 602 is specifically configured to: filter the signal to be identified according to the preset bandwidth; perform smoothing and quantization processing on the filtered signal to obtain a pre-processed step signal.
  • the matching unit 603 is specifically configured to:
  • the difference is smaller than the first preset value, obtain a difference between the preset bandwidth of the to-be-identified signal and the preset bandwidth of the signal template, and if the difference is greater than the second preset value, determine the to-be-determined The identification signal does not match the signal template;
  • the integrated deviation value is smaller than the third preset value, determining that the to-be-identified signal matches the signal template, and obtaining the type and characteristic parameter of the to-be-identified signal.
  • r1 is the ratio of the difference between the power of the 3dB bandwidth of the signal to be identified and the signal template and the power value within the 3dB bandwidth of the signal template
  • r2 is within 10 dB of the signal to be identified and the signal template.
  • r3 is the ratio of the power mean square error of the signal to be identified and the signal template in the 10 dB bandwidth to the power value within the 10 dB bandwidth of the signal template.
  • the components of the apparatus and/or system provided by the embodiments of the present application, and the steps in the method may be concentrated on a single computing device or distributed in multiple calculations. On the network of devices. Alternatively, they may be implemented in program code executable by a computing device. Thus, they may be stored in a storage device by a computing device, or they may be fabricated into individual integrated circuit modules, or a plurality of modules or steps thereof may be implemented as a single integrated circuit module. Thus, the invention is not limited to any specific combination of hardware and software.

Landscapes

  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Image Analysis (AREA)

Abstract

本发明提供了一种基于无线电信号频谱特征模板的信号识别方法，该方法包括：通过提取多种无线电信号的频谱特征，构建信号模板库；根据所述信号模板库中信号模板的表现形式，对待识别信号进行预处理；将预处理后的待识别信号与所述信号频谱模板库进行比较和匹配，获得所述待识别信号的类型和频谱特征参数。本发明还提供了一种基于无线电信号频谱特征模板的信号识别系统，包括模板构建单元、预处理单元及匹配单元。本发明能够通过频谱模板表征各种类的信号，并在保证信号识别准确性的基础上进一步简化了信号识别算法。

Description

基于无线电信号频谱特征模板的信号识别方法及系统
相关申请的交叉引用
本申请要求享有于2014年11月26日提交的名称为“基于无线电信号频谱特征模板的信号识别方法及系统”的中国专利申请CN201410691034.X的优先权，该申请的全部内容通过引用并入本文中。
技术领域
本发明涉及无线通信技术领域，具体涉及基于无线电信号频谱特征模板的信号识别方法及系统。
背景技术
随着无线通信技术的快速发展，大量新型用频系统不断投入使用使得可用频谱资源日益紧缺，大量用频设备工作于相邻频段，易产生相互干扰，使得无线网络电磁空间环境变得日益恶化。伴随着工艺的进步，越来越多的特种设备向着小型化发展，导致大量用频设备集中于狭小区域，使得多种设备之间的干扰频率大大增加。
无线电信号识别技术被广泛应用于干扰源排查领域，可以使干扰排查变得更快速。目前，无线电识别技术主要研究方向是通过提取无线电信号特征(中心频率、带宽、码率、调制方式等)，利用模式识别的方法进行信号识别。现有研究中的信号识别对象都是调制信号，没有考虑到实际更为复杂的无线通信环境对信号产生的影响。如专利“一种信号识别的方法”，利用各类信号功率谱的形态学包络构造了信号模板，模板表述复杂致使信号识别过程在实现过程中比较繁琐，难以应用。
因此，现有信号识别方法是将信号特征进行提取，根据不同信号特征一步一步地排除不相关的信号最终得到信号识别结果，识别过程繁琐。且现有信号识别针对的均为调制信号，如正交相移键控(Quadrature Phase Shift Keying，QPSK)信号、正交振幅调制(Quadrature Amplitude Modulation，QAM)信号等，并未考虑实际信号的其它编码、发送信息等的差异对信号频谱造成的影响。
发明内容
针对现有技术的缺陷，本发明提供一种基于无线电信号频谱特征模板的信号识别方法及系统，能够通过频谱模板表征多种类的信号，并在保证信号识别准确性的基础上进一步简化信号识别算法。
第一方面，本发明提供了一种基于无线电信号频谱特征模板的信号识别方法，所述方法包括：通过提取多种无线电信号的频谱特征，构建信号模板库；根据所述信号模板库中信号模板的表现形式，对待识别信号进行预处理；将预处理后的待识别信号与所述信号模板库的信号模板进行比较和匹配，获得所述待识别信号的类型和特征参数。
优选地，所述通过提取多种无线电信号的频谱特征，构建信号模板库，包括：对多种无线电信号进行滤波、采样和变换，获得频域信号；采用领域平均法对所述频域信号进行平滑处理；对平滑处理后的信号进行量化处理，形成阶梯信号，得到信号模板。
优选地，所述根据所述信号模板库中信号模板的表现形式，对待识别信号进行预处理，包括：根据预设的带宽对待识别的信号进行滤波；对滤波后的信号进行平滑处理及量化处理，获得预处理后的阶梯信号。
优选地，所述将预处理后的待识别信号与所述信号模板库的信号模板进行比较和匹配，获得所述待识别信号的类型和特征参数，包括：获得所述待识别信号的中心频率与所述信号模板库中的各信号模板中心频率的差值，若所述差值大于第一预设值，则判定所述待识别信号与所述信号模板不匹配；若所述差值小于第一预设值，获得所述待识别信号预设带宽与所述信号模板预设带宽的差值，若所述差值大于第二预设值，则判定所述待识别信号与所述信号模板不匹配；若所述差值小于第二预设值，计算所述待识别信号与所述信号模板的综合偏差值，若所述综合偏差值大于第三预设值，则判定所述待识别信号与所述信号模板不匹配；若所述综合偏差值小于第三预设值，则判定所述待识别信号与所述信号模板匹配，并获得所述待识别信号的类型和特征参数。
优选地，所述综合偏差值为：
r＝abs(r1)+abs(r2)+abs(r3)
其中，r为所述综合偏差值，r1为待识别信号与信号模板的3dB带宽内功率的差值与信号模板的3dB带宽内功率值的比值，r2为待识别信号与信号模板的10dB带宽内功率的差值与信号模板的10dB带宽内功率值的比值，r3为待识别信号与信号模板在10dB带宽内的功率均方差与信号模板的10dB带宽内功率值的比值。
第二方面，本发明提供了一种基于无线电信号频谱特征模板的信号识别系统，所述 系统包括：模板构建单元，用于通过提取多种无线电信号的频谱特征，构建信号模板库；预处理单元，用于根据所述信号模板库中信号模板的表现形式，对待识别信号进行预处理；匹配单元，用于将预处理后的待识别信号与所述信号模板库的信号模板进行比较和匹配，获得所述待识别信号的类型和特征参数。
优选地，所述模板构建单元，具体用于：对多种无线电信号进行滤波、采样和变换，获得频域信号；采用领域平均法对所述频域信号进行平滑处理；对平滑处理后的信号进行量化处理，形成阶梯信号，得到信号模板。
优选地，所述预处理单元，具体用于：根据预设的带宽对待识别的信号进行滤波；对滤波后的信号进行平滑处理及量化处理，获得预处理后的阶梯信号。
优选地，所述匹配单元，具体用于：获得所述待识别信号的中心频率与所述信号模板库中的各信号模板中心频率的差值，若所述差值大于第一预设值，则判定所述待识别信号与所述信号模板不匹配；若所述差值小于第一预设值，获得所述待识别信号预设带宽与所述信号模板预设带宽的差值，若所述差值大于第二预设值，则判定所述待识别信号与所述信号模板不匹配；若所述差值小于第二预设值，计算所述待识别信号与所述信号模板的综合偏差值，若所述综合偏差值大于第三预设值，则判定所述待识别信号与所述信号模板不匹配；若所述综合偏差值小于第三预设值，则判定所述待识别信号与所述信号模板匹配，并获得所述待识别信号的类型和特征参数。
优选地，所述综合偏差值为：
r＝abs(r1)+abs(r2)+abs(r3)
其中，r为所述综合偏差值，r1为待识别信号与信号模板的3dB带宽内功率的差值与信号模板的3dB带宽内功率值的比值，r2为待识别信号与信号模板的10dB带宽内功率的差值与信号模板的10dB带宽内功率值的比值，r3为待识别信号与信号模板在10dB带宽内的功率均方差与信号模板的10dB带宽内功率值的比值。
由上述技术方案可知，本发明提供一种基于无线电信号频谱特征模板的信号识别方法及系统，通过提取无线电信号自身的图像特征，利用信号频谱的特定频率、功率点构造频谱模板，来唯一有效地表征不同种类的无线电信号，能够在保证信号识别准确性的基础上进一步简化信号识别算法。
本发明的其它特征和优点将在随后的说明书中阐述，并且，部分地从说明书中变得显而易见，或者通过实施本发明的技术方案而了解。本发明的目的和其他优点可通过在说明书、权利要求书以及附图中所特别指出的结构和/或流程来实现和获得。
附图说明
为了更清楚地说明本发明的实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍。显而易见地，下面描述中的附图仅仅是本发明或者现有技术中的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本发明一实施例提供的基于无线电信号频谱特征模板的信号识别方法的流程示意图。
图2是本发明另一实施例提供的信号模板库构建方法的流程示意图。
图3是本发明另一实施例提供的信号频谱模板示意图。
图4是本发明另一实施例提供的信号频谱模板阶梯选取的示意图。
图5是本发明另一实施例提供的信号识别与匹配的方法的流程示意图。
图6是本发明一实施例提供的基于无线电信号频谱特征模板的信号识别系统的结构示意图。
具体实施方式
下面将结合本发明的实施例中的附图，对本发明的实施例中的技术方案进行清楚、完整地描述。显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
如图1所示，图1示出了本发明一实施例提供的基于无线电信号频谱特征模板的信号识别方法，该方法包括如下步骤：
步骤101、通过提取多种无线电信号的频谱特征，构建信号模板库。
本步骤中，信号模板库并不需要每次都去构建，也可以使用预先构建好的信号模板库。
步骤102、根据所述信号模板库中信号模板的表现形式，对待识别信号进行预处理。
步骤103、将预处理后的待识别信号与所述信号模板库的信号模板进行比较和匹配，获得所述待识别信号的类型和特征参数。
本实施例中，步骤101中信号模板库的构建过程，如图2所示，包括如下步骤：
步骤201、对多种无线电信号进行滤波、采样和变换，获得频域信号。
其中，无线电信号来自于设备发射信号，在经过滤波、采样、变换后得到频域信号。变换时需保证点数相同(如均为1024点FFT)。
步骤202、采用领域平均法对所述频域信号进行平滑处理。
本实施例中，由于无线信道对信号造成的影响，加之噪声的存在使得信号模板在频域的抖动非常剧烈。若用此信号直接绘制信号模板，会导致同种信号的频谱模板在形态学上的相似性大大降低，从而影响信号识别的准确性。为此，需对频域信号进行平滑处理。
本实施例中，采用数字图像处理中的邻域平均法作为信号平滑处理的方法。邻域平均法的平滑滤波是将原图中一个像素的灰度值与它周围邻近像素的灰度值相加，从而降低噪声对图像的影响，如表达式(1)所示。其中，g(i,j)为平滑后的灰度值；S为待处理点邻域点的集合；N为S中点的数量；f(i,j)为图像信号；n(i,j)为噪声信号；F(i,j)则为有噪声的图像信号。在数字图像处理中，邻域通常取待处理点周围的8个点。
Figure PCTCN2015083989-appb-000001
频域信号的幅值曲线可表示为频率的函数，因此邻域点取该频率点周围的两个点即可。本发明中的平滑滤波方法包括以下两种，如表达式(2)和(3)。其中，S’(n)为平滑后的信号幅值；S(n)为原信号幅值。
Figure PCTCN2015083989-appb-000002
Figure PCTCN2015083989-appb-000003
考虑到噪声功率较小时二者平滑性能相差不大，当噪声功率较大时表达式(3)的平滑性能明显优于表达式(2)的方法，因此本发明对频域信号的平滑处理选用表达式(3)的方法。对表达式(3)进一步推导可得表达式(4)。即对于单位为dBm的频谱信号，取均值进行平滑即可。
Figure PCTCN2015083989-appb-000004
如图3所示，当噪声功率较大时，模板频域信号的抖动较大，此时需要做多次平滑处理。对于平滑处理的次数，本发明提出信号最大抖动不超过3dB，即信号为单调的，当某处有抖动时，抖动的波峰与波谷之间幅值之比不大于3dB。
步骤203、对平滑处理后的信号进行量化处理，形成阶梯信号，得到信号模板。
具体来说，经过平滑处理后的模板信号经过量化处理为阶梯信号，即为信号模板。此处量化采用的是非均匀量化。考虑到大部分信号在靠近峰值处的波形变化较缓慢，所以对于阶梯的高度需要小一些，否则简化之后采样点对应频率上间隔就无法保证均匀性；当信号幅度降低时，信号波形进入陡峭段，为了保证频率上的相对均匀，此时的阶梯高度需要选择相对大一些的阶梯；当信号幅度渐渐趋于平缓时，需要将阶梯的高度调整到相对小一些。因此，信号模板构建过程中的阶梯可选择1dB、2dB、3dB、4.5dB、6dB、8dB、10dB、13dB、16dB、19dB、22dB、25dB、28dB、31dB、34dB、37dB、40dB。所有模板库的信号都按照上述量化标准对采样点进行简化构建模板。如图4所示。
经过量化处理的模板信号为阶梯状信号，用一系列频点序列{f1,f2,f3,…,f35}即可将信号模板表示出来，考虑到当信号功率较小时，由于噪声的存在，构建信号模板过程中信号不总能降低到-40dB，此时一些边缘频率(如f1，f35等)的值为空。因此将信号模板构建为如表达式(5)所示。其中，n为FFT变换点数，p1,p2,p3,…,p35是各频率点对应的功率点，P1,P2,P3,…,Pn为功率值。
Figure PCTCN2015083989-appb-000005
当模板信号衰减足够大(即超过40dB)时，35个频率值各不相同，对应的功率点pi＝p36-i；当模板信号衰减不足够大时，以最大衰减-35dB为例，此时f1＝f2＝f3，f33＝f34＝f35，对应的功率点亦相同。这样保证了频谱模板的一致性。
本实施例中，步骤102中的待识别信号的预处理过程如下：
根据预设的带宽对待识别的信号进行滤波；对滤波后的信号进行平滑处理及量化处理，获得预处理后的阶梯信号。
本实施例提出的信号预处理方法，通过对待识别信号进行预处理以使待识别信号与信号模板具有相似的表示形式。因此，信号预处理过程与模板构建过程类似。但是，在模板构建过程中由于信号已知，因此对信号进行滤波时滤波器带宽同样为确定的。当对待识别信号进行滤波时，由于信号类型未知，因此对信号进行滤波的带宽可以定为信号的40dB带宽，以保证采集的待识别信号带宽不会低于其频谱模板所需信号带宽。滤波后的平滑处理及量化与模板构建过程相同。
本实施例中，步骤103中的信号进行识别与匹配的方法，如图5所示，包括如下步骤：
步骤501、获得所述待识别信号的中心频率与所述信号模板库中的各信号模板中心频率的差值。
步骤502、判断步骤501所得的差值是否大于第一预设值，若是，则转至步骤508，否则转至步骤503。
步骤503、获得所述待识别信号预设带宽与所述信号模板预设带宽的差值。
步骤504、判断步骤503所得的差值是否大于第二预设值，若是，则转至步骤508，否则转至步骤505。
步骤505、计算所述待识别信号与所述信号模板的综合偏差值。
步骤506、判断综合偏差值是否大于第三预设值，若是，则转至步骤508，否则转至步骤507。
步骤507、判定所述待识别信号与所述信号模板匹配，并获得所述待识别信号的类型和特征参数。
步骤508、判定所述待识别信号与所述信号模板不匹配。
下面，通过一个具体的实施例来说明信号进行识别与匹配的方法。首先，通过计算获得待识别信号中心频率(即f18)与信号模板中心频率之间的差值。若待识别信号的中心频率偏离信号模板的10dB带宽范围(即f10～f26)则认为该模板与待测信号不匹配。进一步通过比较待识别信号与信号模板的3dB带宽(即f22-f14)之间的差距，若二者之间的差距超过5％，将该信号模板排除在有效信号集之外。
下一步通过计算综合偏差值进行信号的识别与匹配，综合偏差值包括以下三个核心 指标：
1、待识别信号与信号模板的3dB带宽内功率的差值与信号模板的3dB带宽内功率值的比设为r1；在信号模板及待识别信号的表示形式均为
Figure PCTCN2015083989-appb-000006
的情况下，信号的3dB带宽内的功率如表达式(6)所示。
Ρ3＝B3×10-0.3+B2×(10-0.2-10-0.3)+B1×(10-0.1-10-0.2)+B0×(1-10-0.1)
                                                             (6)
＝(f22-f14)×0.5012+(f21-f15)×0.1298+(f20-f16)×0.1634+(f19-f17)×0.2057
利用表达式(6)确定待识别信号及信号模板的3dB带宽内功率，比值r1的计算方法如表达式(7)所示。
Figure PCTCN2015083989-appb-000007
2、待识别信号与信号模板的10dB带宽内功率的差值与信号模板的10dB带宽内功率值的比设为r2；与计算3dB带宽内功率类似，可得到信号的10dB带宽内功率计算方法如表达式(8)所示。
P10＝0.2057×B0+0.1634×B1+0.1298×B2+0.1464×B3+
                                                        (8)
0.1036×B4.5+0.0927×B6+0.0585×B8+0.1×B10
利用表达式(8)可以确定待识别信号及信号模板的3dB带宽的功率值，比值r1的计算方法如表达式(9)所示。
Figure PCTCN2015083989-appb-000008
3、待识别信号与信号模板在10dB带宽内的功率均方差(匹配滤波)与信号模板的10dB带宽内功率值的比设为r3，如表达式(10)所示。
Figure PCTCN2015083989-appb-000009
其中，index为个功率点p在所有功率值P的序列中对应的序号，模板中的i，j，k，m等。index10l和index10r分别为10dB带宽的左右边界，即p10和p26所对应的序号。
定义综合偏差指数r＝abs(r1)+abs(r2)+abs(r3)，计算待识别信号与所有模板信号的综合偏差指数。当待识别信号得的综合偏差指数r最小时所对应的信号模板即为最 相似的信号模板，进一步考察待识别信号与最相似信号模板的相似程度。完成待识别信号与信号模板库的匹配，可以获得信号的频谱特征、信号类型等核心参数。
本实施例提供的基于无线电信号频谱特征模板的信号识别方法，通过提取无线电信号自身的图像特征，利用信号频谱的特定频率、功率点构造频谱模板，来唯一有效地表征不同种类的无线电信号，能够在保证信号识别准确性的基础上进一步简化信号识别算法。
如图6所示，图6示出了本发明一实施例提供的基于无线电信号频谱特征模板的信号识别系统，该系统包括模板构建单元601、预处理单元602以及匹配单元603。
其中，模板构建单元601，用于通过提取多种无线电信号的频谱特征，构建信号模板库；预处理单元602，用于根据所述信号模板库中信号模板的表现形式，对待识别信号进行预处理；匹配单元603，用于将预处理后的待识别信号与所述信号模板库进行比较和匹配，获得所述待识别信号的类型和特征参数。
其中，模板构建单元601，具体用于：对多种无线电信号进行滤波、采样和变换，获得频域信号；采用领域平均法对所述频域信号进行平滑处理；对平滑处理后的信号进行量化处理，形成阶梯信号，得到信号模板。
其中，预处理单元602，具体用于：根据预设的带宽对待识别的信号进行滤波；对滤波后的信号进行平滑处理及量化处理，获得预处理后的阶梯信号。
其中，匹配单元603，具体用于：
获得所述待识别信号的中心频率与所述信号模板库中的各信号模板中心频率的差值，若所述差值大于第一预设值，则判定所述待识别信号与所述信号模板不匹配；
若所述差值小于第一预设值，获得所述待识别信号预设带宽与所述信号模板预设带宽的差值，若所述差值大于第二预设值，则判定所述待识别信号与所述信号模板不匹配；
若所述差值小于第二预设值，计算所述待识别信号与所述信号模板的综合偏差值，若所述综合偏差值大于第三预设值，则判定所述待识别信号与所述信号模板不匹配；
若所述综合偏差值小于第三预设值，则判定所述待识别信号与所述信号模板匹配，并获得所述待识别信号的类型和特征参数。
上述综合偏差值为：
r＝abs(r1)+abs(r2)+abs(r3)
其中，r为所述综合偏差值，r1为待识别信号与信号模板的3dB带宽内功率的差值与信号模板的3dB带宽内功率值的比值，r2为待识别信号与信号模板的10dB带宽内功率的差值与信号模板的10dB带宽内功率值的比值，r3为待识别信号与信号模板在10dB带宽内的功率均方差与信号模板的10dB带宽内功率值的比值。
本领域的技术人员应该明白，上述的本申请实施例所提供的装置和/或系统的各组成部分，以及方法中的各步骤，它们可以集中在单个的计算装置上，或者分布在多个计算装置所组成的网络上。可选地，它们可以用计算装置可执行的程序代码来实现。从而，可以将它们存储在存储装置中由计算装置来执行，或者将它们分别制作成各个集成电路模块，或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。这样，本发明不限制于任何特定的硬件和软件结合。
虽然本发明所揭露的实施方式如上，但所述的内容仅为便于理解本发明技术方案而采用的实施方式，并非用以限定本发明。任何本发明所属领域内的技术人员，在不脱离本发明所揭露的精神和范围的前提下，可以在实施的形式及细节上进行任何的修改与变化，但本发明的专利保护范围，仍须以所附的权利要求书所界定的范围为准。

Claims (10)

  1. 一种基于无线电信号频谱特征模板的信号识别方法，其特征在于，所述方法包括：
    通过提取多种无线电信号的频谱特征，构建信号模板库；
    根据所述信号模板库中信号模板的表现形式，对待识别信号进行预处理；
    将预处理后的待识别信号与所述信号模板库的信号模板进行比较和匹配，获得所述待识别信号的类型和特征参数。
  2. 根据权利要求1所述的方法，其特征在于，所述通过提取多种无线电信号的频谱特征，构建信号模板库，包括：
    对多种无线电信号进行滤波、采样和变换，获得频域信号；
    采用领域平均法对所述频域信号进行平滑处理；
    对平滑处理后的信号进行量化处理，形成阶梯信号，得到信号模板。
  3. 根据权利要求1所述的方法，其特征在于，所述根据所述信号模板库中信号模板的表现形式，对待识别信号进行预处理，包括：
    根据预设的带宽对待识别的信号进行滤波；
    对滤波后的信号进行平滑处理及量化处理，获得预处理后的阶梯信号。
  4. 根据权利要求1所述的方法，其特征在于，所述将预处理后的待识别信号与所述信号模板库的信号模板进行比较和匹配，获得所述待识别信号的类型和特征参数，包括：
    获得所述待识别信号的中心频率与所述信号模板库中的各信号模板中心频率的差值，若所述差值大于第一预设值，则判定所述待识别信号与所述信号模板不匹配；
    若所述差值小于第一预设值，获得所述待识别信号预设带宽与所述信号模板预设带宽的差值，若所述差值大于第二预设值，则判定所述待识别信号与所述信号模板不匹配；
    若所述差值小于第二预设值，计算所述待识别信号与所述信号模板的综合偏差值，若所述综合偏差值大于第三预设值，则判定所述待识别信号与所述信号模板不匹配；
    若所述综合偏差值小于第三预设值，则判定所述待识别信号与所述信号模板匹配，并获得所述待识别信号的类型和特征参数。
  5. 根据权利要求4所述的方法，其特征在于，所述综合偏差值为：
    r＝abs(r1)+abs(r2)+abs(r3)
    其中，r为所述综合偏差值，r1为待识别信号与信号模板的3dB带宽内功率的差值与信号模板的3dB带宽内功率值的比值，r2为待识别信号与信号模板的10dB带宽内功率的差值与信号模板的10dB带宽内功率值的比值，r3为待识别信号与信号模板在10dB带宽内的功率均方差与信号模板的10dB带宽内功率值的比值。
  6. 一种基于无线电信号频谱特征模板的信号识别系统，其特征在于，所述系统包括：
    模板构建单元，用于通过提取多种无线电信号的频谱特征，构建信号模板库；
    预处理单元，用于根据所述信号模板库中信号模板的表现形式，对待识别信号进行预处理；
    匹配单元，用于将预处理后的待识别信号与所述信号模板库的信号模板进行比较和匹配，获得所述待识别信号的类型和特征参数。
  7. 根据权利要求6所述的系统，其特征在于：
    所述模板构建单元被设置为对多种无线电信号进行滤波、采样和变换，获得频域信号；采用领域平均法对所述频域信号进行平滑处理；对平滑处理后的信号进行量化处理，形成阶梯信号，得到信号模板。
  8. 根据权利要求6所述的系统，其特征在于：
    所述预处理单元被设置为根据预设的带宽对待识别的信号进行滤波；对滤波后的信号进行平滑处理及量化处理，获得预处理后的阶梯信号。
  9. 根据权利要求6所述的系统，其特征在于：
    所述匹配单元被设置为获得所述待识别信号的中心频率与所述信号模板库中的各信号模板中心频率的差值，若所述差值大于第一预设值，则判定所述待识别信号与所述信号模板不匹配；若所述差值小于第一预设值，获得所述待识别信号预设带宽与所述信号模板预设带宽的差值，若所述差值大于第二预设值，则判定所述待识别信号与所述信号模板不匹配；若所述差值小于第二预设值，计算所述待识别信号与所述信号模板的综合偏差值，若所述综合偏差值大于第三预设值，则判定所述待识别信号与所述信号模板不匹配；若所述综合偏差值小于第三预设值，则判定所述待识别信号与所述信号模板匹配，并获得所述待识别信号的类型和特征参数。
  10. 根据权利要求9所述的系统，其特征在于，所述综合偏差值为：
    r＝abs(r1)+abs(r2)+abs(r3)
    其中，r为所述综合偏差值，r1为待识别信号与信号模板的3dB带宽内功率的差值与信号模板的3dB带宽内功率值的比值，r2为待识别信号与信号模板的10dB带宽内功率的差值与信号模板的10dB带宽内功率值的比值，r3为待识别信号与信号模板在10dB带宽内的功率均方差与信号模板的10dB带宽内功率值的比值。
PCT/CN2015/083989 2014-11-26 2015-07-14 基于无线电信号频谱特征模板的信号识别方法及系统 WO2016082562A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410691034.X 2014-11-26
CN201410691034.XA CN104468001B (zh) 2014-11-26 2014-11-26 基于无线电信号频谱特征模板的信号识别方法及系统

Publications (1)

Publication Number Publication Date
WO2016082562A1 true WO2016082562A1 (zh) 2016-06-02

Family

ID=52913452

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/083989 WO2016082562A1 (zh) 2014-11-26 2015-07-14 基于无线电信号频谱特征模板的信号识别方法及系统

Country Status (2)

Country Link
CN (1) CN104468001B (zh)
WO (1) WO2016082562A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112818596A (zh) * 2021-01-29 2021-05-18 广东振子电子科技有限公司 一种智能干扰识别定位方法及其系统
CN113138386A (zh) * 2020-01-16 2021-07-20 为昇科科技股份有限公司 生命体雷达检测系统及其判断方法、特征数据库建立方法

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104468001B (zh) * 2014-11-26 2017-04-19 北京邮电大学 基于无线电信号频谱特征模板的信号识别方法及系统
CN105142145B (zh) * 2015-08-21 2019-02-12 中国电子科技集团公司第二十九研究所 一种基于前沿特征的Wi-Fi设备识别方法
CN106452716B (zh) * 2016-11-15 2019-09-06 厦门大学 基于Hash指纹的无人机信号识别检测方法
CN107544943B (zh) * 2017-08-28 2021-02-19 湖南时变通讯科技有限公司 一种基于fft库函数的部分频点的计算方法及装置
CN108154083B (zh) * 2017-12-04 2020-06-30 厦门市美亚柏科信息股份有限公司 一种基于svm的信号识别方法及终端
CN108469602B (zh) * 2018-03-29 2021-09-07 东南大学 一种基于谱特征提取的脉冲信号类型自动判别方法
CN108768562B (zh) * 2018-05-07 2020-10-27 西北工业大学 一种基于遥控与图传信号的非法无人机识别方法
CN109104257B (zh) * 2018-07-04 2020-08-21 北京邮电大学 一种无线信号检测方法及装置
CN114330409A (zh) * 2020-09-24 2022-04-12 广州慧睿思通科技股份有限公司 信号处理方法、电子设备和存储介质
CN113517944A (zh) * 2021-04-09 2021-10-19 广州辰创科技发展有限公司 一种无线电压制方法、系统及存储介质
CN116015497A (zh) * 2022-12-27 2023-04-25 成都爱科特科技发展有限公司 一种信号频谱模板匹配方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080215322A1 (en) * 2004-02-18 2008-09-04 Koninklijke Philips Electronic, N.V. Method and System for Generating Training Data for an Automatic Speech Recogniser
CN102780534A (zh) * 2012-07-04 2012-11-14 中国电子科技集团公司第四十一研究所 一种通用的电磁信号自动搜索方法
CN103929260A (zh) * 2014-03-12 2014-07-16 上海风格信息技术股份有限公司 一种无线广播异常频点的分析方法
CN104468001A (zh) * 2014-11-26 2015-03-25 北京邮电大学 基于无线电信号频谱特征模板的信号识别方法及系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103117061B (zh) * 2013-02-05 2016-01-20 广东欧珀移动通信有限公司 一种基于语音的动物识别方法及装置
CN104158611B (zh) * 2014-08-20 2016-10-26 西安电子科技大学 基于频谱分析的无线信号干扰检测系统及方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080215322A1 (en) * 2004-02-18 2008-09-04 Koninklijke Philips Electronic, N.V. Method and System for Generating Training Data for an Automatic Speech Recogniser
CN102780534A (zh) * 2012-07-04 2012-11-14 中国电子科技集团公司第四十一研究所 一种通用的电磁信号自动搜索方法
CN103929260A (zh) * 2014-03-12 2014-07-16 上海风格信息技术股份有限公司 一种无线广播异常频点的分析方法
CN104468001A (zh) * 2014-11-26 2015-03-25 北京邮电大学 基于无线电信号频谱特征模板的信号识别方法及系统

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113138386A (zh) * 2020-01-16 2021-07-20 为昇科科技股份有限公司 生命体雷达检测系统及其判断方法、特征数据库建立方法
CN112818596A (zh) * 2021-01-29 2021-05-18 广东振子电子科技有限公司 一种智能干扰识别定位方法及其系统

Also Published As

Publication number Publication date
CN104468001B (zh) 2017-04-19
CN104468001A (zh) 2015-03-25

Similar Documents

Publication Publication Date Title
WO2016082562A1 (zh) 基于无线电信号频谱特征模板的信号识别方法及系统
US10594529B1 (en) Variational design of companders for PAPR reduction in OFDM systems
CN108764077B (zh) 一种基于卷积神经网络的数字信号调制分类方法
CN111614398B (zh) 基于异或神经网络的调制格式及信噪比识别方法及装置
CN114268526B (zh) 基于图的度特征的bpsk及qpsk信号调制识别方法
Wang et al. Deep learning based modulation recognition with multi-cue fusion
CN114025379B (zh) 一种宽带多信号检测方法、装置和设备
CN112749633B (zh) 分离与重构的个体辐射源识别方法
CN110045322B (zh) 一种基于高分辨测向语图智能识别的短波自动测向方法
CN115146670A (zh) 基于数据增强和对比学习的射频指纹识别方法及系统
CN110135390B (zh) 基于主信号抑制的辐射源个体识别方法
CN112034434A (zh) 基于稀疏时频检测卷积神经网络的雷达辐射源识别方法
Wang et al. Specific emitter identification based on deep adversarial domain adaptation
Zhang et al. Specific emitter identification based on radio frequency fingerprint using multi-scale network
CN103856276B (zh) 频谱非平坦噪声基底实时自动校正方法
CN110048977B (zh) 基于灰度共生矩阵纹理特征检测的短波信号体制识别方法及装置
CN101098557B (zh) 频域中的信噪比确定
CN114006798A (zh) 一种信号处理方法、装置、电子设备及存储介质
CN111814703B (zh) 一种非重构条件下基于hb的信号联合特征提取方法
CN114501329A (zh) 一种基于1d-cnn的室内定位模型的环境自适应训练方法及装置
CN109450835B (zh) 基于hsca提取通信信号特征的调制识别方法
KR100957073B1 (ko) 지문영상의 품질에 기초한 선택적 영상압축장치 및 방법, 그리고 이를 이용한 인증장치
US8892052B2 (en) Methods for determining whether a signal includes a wanted signal and apparatuses configured to determine whether a signal includes a wanted signal
Zhao et al. Efficient and Appropriate Key Generation Scheme in Different IoT Scenarios
CN114997299B (zh) 一种资源受限环境下的射频指纹识别方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15862921

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 19/10/2017)

122 Ep: pct application non-entry in european phase

Ref document number: 15862921

Country of ref document: EP

Kind code of ref document: A1