WO2022217641A1 - 基于自适应滤波器系数的射频指纹统一表达方法及电子设备 - Google Patents

基于自适应滤波器系数的射频指纹统一表达方法及电子设备 Download PDF

Info

Publication number
WO2022217641A1
WO2022217641A1 PCT/CN2021/089536 CN2021089536W WO2022217641A1 WO 2022217641 A1 WO2022217641 A1 WO 2022217641A1 CN 2021089536 W CN2021089536 W CN 2021089536W WO 2022217641 A1 WO2022217641 A1 WO 2022217641A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
fingerprint
radio frequency
baseband signal
coefficients
Prior art date
Application number
PCT/CN2021/089536
Other languages
English (en)
French (fr)
Inventor
胡爱群
俞佳宝
Original Assignee
网络通信与安全紫金山实验室
东南大学
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 网络通信与安全紫金山实验室, 东南大学 filed Critical 网络通信与安全紫金山实验室
Publication of WO2022217641A1 publication Critical patent/WO2022217641A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
    • H03H21/00—Adaptive networks

Definitions

  • the invention relates to the field of information security, in particular to a method for uniformly expressing radio frequency fingerprints based on adaptive filter coefficients and an electronic device.
  • the radio frequency fingerprint originates from the difference of the transmitter circuit design and the manufacturing tolerance of the hardware circuit in the production process, and the device parameters of different production batches or even the same production batch have subtle differences. In addition, the differences between transmitters designed by different manufacturers are even more pronounced due to differences in component types, component layouts, and PCB routing.
  • the RF fingerprints of devices are unique and difficult to clone, so they can be used for identification and authentication of wireless devices.
  • the existing radio frequency fingerprint identification technology can accurately distinguish devices using the same frequency, the same bandwidth and the same modulation method, and has very good practical value.
  • the existing disclosed RF fingerprint technologies are completely different for the RF fingerprint extraction methods of the same transmitter under different parameter configurations (such as different modulation methods, different symbol rates, etc.), and most of the technologies also require the existence of the same data sequence. What is data-related RF fingerprints, these methods do not have the uniformity of RF fingerprints.
  • the RF fingerprint is essentially determined by the hardware circuit defects of the transmitter, in the same frequency band, the hardware circuits passed by the same transmitter with different modulation methods and different symbol rates are consistent. Therefore, in theory, it is possible to configure these different When parameters are used, a unified method is used to extract a unified RF fingerprint.
  • the completely different RF fingerprint expressions under different parameter configurations do not utilize the establishment and use of RF fingerprint expression standards. In actual use, the existing method needs to traverse all the signals under the configuration of parameters to extract the radio frequency fingerprint for storage, which is poor in practicability for equipment with variable parameters. Therefore, there is an urgent need for a simple and effective method for the unified expression of RF fingerprints.
  • the present invention provides a unified expression method and electronic device for radio frequency fingerprints based on adaptive filter coefficients.
  • the ideal baseband signal of the fingerprint is used as the filter input, and the RF fingerprint, which is uniformly expressed as N-order adaptive filter coefficients, is estimated by the adaptive algorithm, and finally N+1 filter coefficients are extracted as the RF fingerprint to realize the different modulation methods, Devices with different symbol rates use the same RF fingerprint extraction purpose.
  • a method for uniformly expressing radio frequency fingerprints based on adaptive filter coefficients comprising the following steps:
  • the receiver receives the radio frequency signal r(t) at the sampling rate f s , obtains the discrete signal r(n), and obtains the entire baseband signal y r (n) through down-conversion;
  • the RF fingerprint is uniformly expressed as an N-order filter RFF(z) in the z domain, and the coefficients of the N-order filter RFF(z) are adjusted by an adaptive algorithm, and the convergence is When the error signal e(n) between the filter output signal y f (n) and the baseband signal y(n) is minimized, represents the adaptive filter after the convergence of the N-order filter RFF(z);
  • the error signal e(n) is specifically expressed as:
  • N represents the order of the N-order filter
  • n represents the discrete time
  • k represents the coefficient index of the N-order filter
  • the radio frequency signal in the step (1) is a wireless signal
  • the receiver sampling rate f s is greater than the Nyquist sampling rate, that is, oversampling is required.
  • the carrier frequency offset and phase offset estimation methods in the step (2) include but are not limited to the preamble complex correlation method or the difference method.
  • performing symbol demodulation and re-modulation on the baseband signal y(n) containing the RF fingerprint in the step (3) is to process a complete frame of signal, including the preamble sequence and the random data sequence.
  • the adaptive algorithm used when adjusting the coefficient of RFF(z) is any one of the zero-forcing algorithm, the steepest descent method, the LMS algorithm, and the RLS algorithm.
  • the mean square error of e(n) is minimized, and the final coefficient of the adaptive filter is obtained as:
  • An electronic device comprising:
  • the signal receiving module is used to receive the radio frequency signal and perform sampling and down-conversion processing to obtain the baseband signal;
  • the first processing module is used to perform signal synchronization, carrier frequency offset and phase offset estimation on the baseband signal to obtain the baseband signal including the radio frequency fingerprint;
  • the second processing module is used to demodulate and re-modulate the baseband signal containing the radio frequency fingerprint to obtain an ideal baseband signal without the radio frequency fingerprint;
  • the adaptive calculation module uses an adaptive algorithm to adjust the coefficients of the N-order filter used to express the RF fingerprint.
  • the adjustment result converges so that the error signal between the output signal of the N-order filter and the baseband signal containing the RF fingerprint is the smallest.
  • the input signal of the filter is an ideal baseband signal;
  • the RF fingerprint extraction module is used to extract the N+1 coefficients of the converged adaptive filter as the final RF fingerprint.
  • the RF fingerprint is uniformly expressed as an N-order filter RFF(z) in the z domain, and the expression of the Nth-order filter RFF(z) is:
  • the error signal e(n) is specifically expressed as:
  • the final RF fingerprint expression is:
  • N represents the order of the N-order filter
  • n represents the discrete time
  • k represents the coefficient index of the N-order filter
  • the received radio frequency signal is a wireless signal
  • the first processing module uses the local preamble signal and the received baseband signal to perform complex correlation and differential processing to achieve synchronization, carrier frequency offset and phase offset estimate.
  • the present invention is based on the uniform expression method of radio frequency fingerprints based on adaptive filter coefficients, which can provide the same radio frequency fingerprint representation for transmitters of different types and different modulation modes, the extracted radio frequency fingerprints are simple and effective, the complexity of the radio frequency fingerprint system is reduced, and the It can provide technical means for establishing a unified radio frequency fingerprint expression standard; it can be applied to various wireless communication systems, especially narrow-band wireless communication systems.
  • Fig. 1 is the flow chart of the radio frequency fingerprint unified expression method based on adaptive filter coefficient of the present invention
  • Fig. 2 is the frame format of QPSK signal and 16-QAM signal in the embodiment of the present invention
  • 3 is a comparison diagram of the initial part of the signal segment of the baseband signal containing the radio frequency fingerprint and 30dB noise in the QPSK signal and the ideal signal after re-modulation in the embodiment of the present invention
  • 4 is a comparison diagram of the initial part of the signal segment of the baseband signal containing the radio frequency fingerprint and 30dB noise in the 16-QAM signal and the ideal signal after re-modulation in the embodiment of the present invention
  • FIG. 5 is a comparison diagram of the real part of the baseband received signal corresponding to the QPSK signal in the embodiment of the present invention, the real part of the output signal of the adaptive filter, and the errors of the two;
  • FIG. 6 is a comparison diagram of the real part of the baseband received signal corresponding to the 16-QAM signal, the real part of the output signal of the adaptive filter, and the errors of the two according to the embodiment of the present invention.
  • This embodiment provides a uniform expression method for radio frequency fingerprints based on adaptive filter coefficients, as shown in FIG. 1 , including the following steps:
  • the receiver receives the radio frequency signal r(t) at the sampling rate f s , obtains the discrete signal r(n), and obtains all baseband signals y r (n) through down-conversion, t refers to continuous time, and n refers to discrete time.
  • QPSK and QAM modulation techniques commonly used in wireless digital communication systems are selected, and the QPSK signal with a 2MHz symbol rate and a 16-QAM signal with a 500kHz symbol rate are collected and down-converted at a sampling rate of 10Msps, respectively, to obtain the corresponding baseband signal and Therefore, the receiver oversamples these two signals by a factor of 5 and 20, respectively.
  • the selected QPSK and 16-QAM signal frame formats are shown in FIG. 2 , and both include a preamble part of 8 all-zero symbols and random data of 248 symbols. Since high oversampling has been adopted during reception, signal interpolation is not performed to increase the sampling rate.
  • Use the local preamble signal and the received signal to perform complex correlation and differential processing to achieve synchronization, carrier frequency offset and phase offset estimation, and then remove the residual frequency offset and phase offset from the baseband QPSK and 16-QAM signals containing 256 symbols , the complete baseband signals y QPSK (n) and y QAM (n) containing the RF fingerprint are obtained.
  • symbol demodulation is performed on the complete baseband signals y QPSK (n) and y QAM (n) containing RF fingerprints, respectively, to obtain the corresponding symbols s QPSK (n) and s QAM (n), and then the The corresponding modulation produces ideal signals x QPSK (n) and x QAM (n) without RF fingerprints.
  • the two kinds of signals selected in this embodiment are complex baseband signals, which include two-channel signals of real part and imaginary part. The comparison of the real part of the signals before symbol demodulation and after re-modulation is shown in FIG. 3 and FIG. 4 , respectively. Re-modulation is relative to the first modulation on the transmitter side. At the receiver side, after symbol demodulation of the baseband signal y(n), it is modulated again, and the corresponding demodulated QPSK and 16-QAM signals are re-modulated. QPSK and 16-QAM modulation.
  • the RF fingerprint is uniformly expressed as an N-order filter in the z domain:
  • RFF(z) The coefficient of RFF(z) is adjusted by an adaptive algorithm, and the convergence is When the error e(n) between the filter output signal yf (n) and the baseband signal y(n) is minimized.
  • RFF k denotes the k-th order coefficient of the N-order filter.
  • a 4th-order filter is selected to express the RF fingerprint uniformly, then the error signal e(n) is:
  • the coefficient of RFF(z) can be adjusted by adopting zero-forcing algorithm, steepest descent method, LMS algorithm or RLS algorithm.
  • the zero-forcing algorithm is an adjustment algorithm with the minimum peak distortion as the criterion.
  • the parameters of the adaptive filter are automatically adjusted to eliminate the inter-symbol interference as much as possible to achieve the best equalization effect.
  • the steepest descent method is a gradient algorithm that can introduce functions and parameter vectors including basic parameters such as signal amplitude, angular frequency and phase, and then use the steepest descent method to estimate and track changes in parameters.
  • the obtained adaptive filter has Stronger robustness and faster convergence speed.
  • the LMS algorithm is a least mean square algorithm based on the minimum mean square error MMSE criterion.
  • the RLS algorithm is a recursive least squares algorithm, which replaces the least mean square criterion with the time average criterion of the square, and performs iterative calculation according to time, that is, averages the squares of all errors from the starting time to the current time , and minimize, the RLS algorithm has fast tracking ability and can be applied to time-varying channels.
  • the LMS algorithm is used as an example for description:
  • the output of the adaptive filter is basically consistent with the received baseband signal, and the real part error It is very small, and the imaginary part behaves similarly, so it is not repeated. Therefore, the adaptive filter can give a good characterization of the RF fingerprint of the device.
  • the method to minimize the error signal e(n) also includes the least squares (LS) criterion, the maximum signal-to-noise ratio criterion or the statistical detection criterion.
  • the method of the present invention is used to extract the radio frequency fingerprints of QPSK and 16-QAM signals with different symbol rates transmitted by the same device under a signal-to-noise ratio of 30dB, as shown in Table 1.
  • the present invention provides a simple and effective uniform expression method for the radio frequency fingerprints of wireless devices, and the radio frequency fingerprints extracted under different symbol rates and different modulation modes are basically consistent.

Landscapes

  • Digital Transmission Methods That Use Modulated Carrier Waves (AREA)

Abstract

一种基于自适应滤波器系数的射频指纹统一表达方法及电子设备，其中方法包括：接收机过采样接收无线信号，通过下变频得到全部基带信号；进行信号同步、载波频偏和相偏估计，去除残留频偏与相偏，得到包含射频指纹的基带信号；对基带信号进行符号解调和重新调制，得到不含射频指纹的理想基带信号；通过自适应算法对统一表达为N阶自适应滤波器系数的射频指纹进行估计；最终提取N+1个滤波器系数作为射频指纹。

Description

基于自适应滤波器系数的射频指纹统一表达方法及电子设备 技术领域
本发明涉及信息安全领域，尤其涉及一种基于自适应滤波器系数的射频指纹统一表达方法及电子设备。
背景技术
射频指纹(RFF)源自发射机电路设计的差异和生产过程中硬件电路的制造容差，不同生产批次甚至同一生产批次的器件参数都有着细微差别。此外，由于其元件类型、元件布局以及PCB走线等存在差异，不同厂家设计的发射机之间的差异性更加显著。
设备的射频指纹都具有唯一性和难以克隆性，从而可以被用于无线设备的身份识别与认证。具体来说，现有的射频指纹识别技术可以准确区别即使采用了相同频率、相同带宽和相同调制方式的设备，具有非常好的实用价值。然而，现有公开的射频指纹技术对于同一发射机在不同参数配置下(例如不同调制方式、不同符号速率等)的射频指纹提取方法截然不同，且大部分技术还要求存在相同的数据序列，提取的是数据相关的射频指纹，这些方法都不具备射频指纹统一性。
由于射频指纹实质上是由发射机的硬件电路缺陷决定的，在同一频带内，同一个发射机采用不同调制方式和不同符号速率时通过的硬件电路保持一致，因此，理论上可以在配置这些不同参数时使用统一的方法提取出统一的射频指纹。此外，不同参数配置下截然不同的射频指纹表达不利用射频指纹表达标准的建立和使用。在实际使用时，现有方法需要遍历所有参数配置下的信号提取射频指纹进行入库，对参数可变的设备实用性差。所以迫切需要一种简单有效的射频指纹统一表达方法。
发明内容
发明目的：为了解决现有技术存在的问题，本发明提供了一种基于自适应滤波器系数的射频指纹统一表达方法及电子设备，通过将射频指纹建模为N阶滤波器，将不含射频指纹的理想基带信号作为滤波器输入，通过自适应算法对统一表达为N阶自适应滤波器系数的射频指纹进行估计，最终提取N+1个滤波器系数作为射频指纹，实现对不同调制方式、不同符号速率的设备采用相同的射频指纹提取的目的。
技术方案：为实现上述技术目的，本发明采用了如下技术方案：
一种基于自适应滤波器系数的射频指纹统一表达方法，其特征在于，包括以下步骤：
(1)接收机以采样率f s接收射频信号r(t)，得到离散信号r(n)，通过下变频得到全部基带信号y r(n)；
(2)进行信号同步、载波频偏和相偏估计，去除基带信号y r(n)中的残留频偏和相偏，得到包含射频指纹的基带信号y(n)；
(3)对包含射频指纹的基带信号y(n)进行符号解调和重新调制，得到不含射频指纹的理想基带信号x(n)；
(4)将射频指纹统一表达为z域的N阶滤波器RFF(z)，通过自适应算法对N阶滤波器RFF(z)的系数进行调整，收敛为
Figure PCTCN2021089536-appb-000001
时使得滤波器输出信号y f(n)和基带信号y(n)的误差信号e(n)最小，
Figure PCTCN2021089536-appb-000002
表示N阶滤波器RFF(z)收敛后的自适应滤波器；
(5)提取收敛后的自适应滤波器
Figure PCTCN2021089536-appb-000003
的N+1个系数作为最终的射频指纹。
具体地，所述步骤(4)中，N阶滤波器RFF(z)的表达式为：
Figure PCTCN2021089536-appb-000004
误差信号e(n)具体表达为：
Figure PCTCN2021089536-appb-000005
步骤(5)提取到的收敛后的自适应滤波器
Figure PCTCN2021089536-appb-000006
的N+1个系数，作为最终的射频指纹表达式为：
Figure PCTCN2021089536-appb-000007
其中，N表示N阶滤波器的阶数，n表示离散时间，k表示N阶滤波器的系数索引。
具体地，所述步骤(1)中的射频信号为无线信号，且接收机采样率f s大于奈奎斯特采样率，即要求过采样。
具体地，所述步骤(2)中的载波频偏和相偏估计方法包括但不限于前导复相关法或差分法。
具体地，所述步骤(3)中对包含射频指纹的基带信号y(n)进行符号解调和重新调制，是对完整的一帧信号进行处理，包括前导序列和随机数据序列。
具体地，所述步骤(4)中，对RFF(z)的系数进行调整时采用的自适应算法为迫零算法、最速下降法、LMS算法、RLS算法中的任一种。
具体地，所述步骤(4)中，使用LMS算法对RFF(z)的系数进行调整时，使得e(n)的均方误差最小，得到自适应滤波器的最终系数为：
Figure PCTCN2021089536-appb-000008
Figure PCTCN2021089536-appb-000009
是使e(n)e *(n)具有最小值的RFF(z)的值的集合，e *(n)表示误差信号e(n)的复共轭。
一种电子设备，其特征在于，包括：
信号接收模块，用于接收射频信号并进行采样和下变频处理，获得基带信号；
第一处理模块，用于对基带信号进行信号同步、载波频偏和相偏估计，得到包含射频指纹的基带信号；
第二处理模块，用于对包含射频指纹的基带信号进行符号解调和重新调制，得到不含射频指纹的理想基带信号；
自适应计算模块，采用自适应算法对用于表达射频指纹的N阶滤波器的系数进行调整，调整结果收敛使得N阶滤波器的输出信号和包含射频指纹的基带信号的误差信号最小，N阶滤波器的输入信号为理想基带信号；
射频指纹提取模块，用于提取出收敛后的自适应滤波器的N+1个系数作为最终的射频指纹。
具体地，所述自适应计算模块中，将射频指纹统一表达为z域的N阶滤波器RFF(z)，第N阶滤波器RFF(z)的表达式为：
Figure PCTCN2021089536-appb-000010
误差信号e(n)具体表达为：
Figure PCTCN2021089536-appb-000011
射频指纹提取模块中，最终的射频指纹表达式为：
Figure PCTCN2021089536-appb-000012
其中，N表示N阶滤波器的阶数，n表示离散时间，k表示N阶滤波器的系数索引。
具体地，所述信号接收模块中，接收的射频信号为无线信号，第一处理模块使用本地的前导符信号和接收的基带信号信号进行复相关和差分处理，实现同步、载波频偏和相偏估计。
有益效果：本发明与现有技术相比，其显著优点如下：
本发明基于自适应滤波器系数的射频指纹统一表达方法，可以为不同类型、不同调制方式的发射机提供相同的射频指纹表征，提取的射频指纹简单有效，降低了射频指纹系统的复杂度，进而可为建立统一的射频指纹表达标准提供技术手段；能够适用于各种无线通信体制，特别是窄带无线通信体制。
附图说明
图1为本发明的基于自适应滤波器系数的射频指纹统一表达方法的流程图；
图2为本发明实施例中QPSK信号和16-QAM信号的帧格式；
图3为本发明实施例中QPSK信号含射频指纹和30dB噪声的基带信号与重新调制后理想信号的开始部分信号段对比图；
图4为本发明实施例中16-QAM信号含射频指纹和30dB噪声的基带信号与重新调制后理想信号的开始部分信号段对比图；
图5为本发明实施例中QPSK信号对应的基带接收信号实部、自适应滤波器输出信号实部以及两者误差对比图；
图6为本发明实施例中16-QAM信号对应的基带接收信号实部、自适应滤波器输出信号实部以及两者误差对比图。
具体实施方式
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅用于解释本发明，并不用于限定本发明。
本实施例提供一种基于自适应滤波器系数的射频指纹统一表达方法，如图1 所示，包括以下步骤：
(1)接收机以采样率f s接收射频信号r(t)，得到离散信号r(n)，通过下变频得到全部基带信号y r(n)，t指连续时间，n表示离散时间。
本实施例中，选用了无线数字通信系统中普遍使用的QPSK和QAM调制技术，以10Msps采样率分别对2MHz符号速率的QPSK信号和500kHz符号速率的16-QAM信号进行采集和下变频，得到对应的基带信号
Figure PCTCN2021089536-appb-000013
和
Figure PCTCN2021089536-appb-000014
因此，接收端对这两个信号分别进行了5倍和20倍过采样。
(2)进行信号同步、载波频偏和相偏估计，去除基带信号y r(n)中的残留频偏和相偏，得到包含射频指纹的基带信号y(n)。
本实施例中，选用的QPSK和16-QAM信号帧格式如图2所示，均包含8个全零符号的前导部分和248个符号的随机数据。由于接收时已经采用了高倍过采样，不进行信号内插提高采样率。使用本地的前导符信号和接收的信号进行复相关和差分等处理，实现同步、载波频偏和相偏估计，然后对包含256个符号的基带QPSK和16-QAM信号去除残留频偏和相偏，得到包含射频指纹的完整基带信号y QPSK(n)和y QAM(n)。
(3)对基带信号y(n)进行符号解调和重新调制，得到不含射频指纹的理想基带信号x(n)。
本实施例中，对包含射频指纹的完整基带信号y QPSK(n)和y QAM(n)分别进行符号解调，得到对应的符号s QPSK(n)和s QAM(n)，然后对其进行相应的调制生成不含射频指纹的理想信号x QPSK(n)和x QAM(n)。本实施例选用的两种信号为复基带信号，其包括实部和虚部两路信号，符号解调前和重新调制后信号的实部对比分别如图3和图4所示。重新调制是相对于发射机端的第一次调制而言，在接收机端，对基带信号y(n)进行符号解调后，再次调制，对应的解调后的QPSK和16-QAM信号重新进行QPSK和16-QAM调制。
(4)将射频指纹统一表达为z域的N阶滤波器：
Figure PCTCN2021089536-appb-000015
通过自适应算法对RFF(z)的系数进行调整，收敛为
Figure PCTCN2021089536-appb-000016
时使得滤波器输出信号y f(n)和基带信号y(n)的误差e(n)最小。RFF k表示N阶滤波器的第k阶系数。
本实施例中，选用4阶滤波器统一表达射频指纹，则误差信号e(n)为：
Figure PCTCN2021089536-appb-000017
本发明中对RFF(z)的系数进行调整，可采用迫零算法、最速下降法、LMS算法或RLS算法等。迫零算法是一种以最小峰值畸变为准则的调整算法，根据给定的输入信号，自动调整自适应滤波器的参数，尽量消除码间干扰，使均衡效果达到最佳。最速下降法，是一种梯度算法，可引入包括信号幅值、角频率和相位等基本参数的函数和参数矢量，然后采用最速下降法进行估计，跟踪参数的变化，获得的自适应滤波器具有较强的鲁棒性和较快的收敛速度。LMS算法，是一种基于最小均方误差MMSE准则的最小均方算法。RLS算法，是一种递推最小二乘法算法，其用二乘方的时间平均准则取代最小均方准则，并按照时间进行迭代计算，即对从起始时刻到当前时刻所有误差的平方进行平均，并最小化，RLS算法具有快速的跟踪能力，能够适用于时变信道中。
本实施例中，以LMS算法为例进行说明：
使用LMS自适应算法对RFF(z)的系数进行调整，使得e(n)的均方误差最小，得到自适应滤波器的最终系数：
Figure PCTCN2021089536-appb-000018
Figure PCTCN2021089536-appb-000019
是使e(n)e *(n)具有最小值的RFF(z)的值的集合，e *(n)表示误差信号e(n)的复共轭，LMS算法通过期望信号y(n)与输出信号y f(n)之差e(n)来自动调节自适应滤波器的参数，使下一时刻的输出y(n+1)能够更加接近期望信号。根据实际情况改变LMS滤波器初始权重和步长μ以达到最优的收敛情况和结果。本实施例两种信号对应的自适应滤波器输出和接收基带信号实部部分的对比如图5和图6所示，可以看到，自适应滤波器输出和接收基带信号基本吻合，实部误 差很小，虚部表现类似，没有赘述。因此，自适应滤波器可以对设备的射频指纹进行良好的表征。使得误差信号e(n)最小的方法除采用最小均方误差(LMS)准则外，还包括最小二乘(LS)准则、最大信噪比准则或统计检测准则等。
(5)提取收敛后的自适应滤波器
Figure PCTCN2021089536-appb-000020
的N+1个系数作为最终的射频指纹：
Figure PCTCN2021089536-appb-000021
本实施例中，使用本发明方法对30dB信噪比下，同一设备发射的不同符号速率的QPSK和16-QAM两种信号提取的射频指纹如表1所示。
表1
Figure PCTCN2021089536-appb-000022
可以看到，本发明对无线设备的射频指纹提供了一种简单有效的统一性表达方法，在不同符号速率、不同调制方式时提取的射频指纹基本保持一致。
以上所述仅为本发明的一种较佳实施例而已，不能以此来限定本发明之权利范围，因此依本发明权利要求所作的等同变化，仍属本发明所涵盖的范围。

Claims (10)

  1. 一种基于自适应滤波器系数的射频指纹统一表达方法，其特征在于，包括以下步骤：
    (1)接收机以采样率f s接收射频信号r(t)，得到离散信号r(n)，通过下变频得到全部基带信号y r(n)；
    (2)进行信号同步、载波频偏和相偏估计，去除基带信号y r(n)中的残留频偏和相偏，得到包含射频指纹的基带信号y(n)；
    (3)对包含射频指纹的基带信号y(n)进行符号解调和重新调制，得到不含射频指纹的理想基带信号x(n)；
    (4)将射频指纹统一表达为z域的N阶滤波器RFF(z)，通过自适应算法对N阶滤波器RFF(z)的系数进行调整，收敛为
    Figure PCTCN2021089536-appb-100001
    时使得滤波器输出信号y f(n)和基带信号y(n)的误差信号e(n)最小，
    Figure PCTCN2021089536-appb-100002
    表示N阶滤波器RFF(z)收敛后的自适应滤波器；
    (5)提取收敛后的自适应滤波器
    Figure PCTCN2021089536-appb-100003
    的N+1个系数作为最终的射频指纹。
  2. 根据权利要求1所述的一种基于自适应滤波器系数的射频指纹统一表达方法，其特征在于，所述步骤(4)中，N阶滤波器RFF(z)的表达式为：
    Figure PCTCN2021089536-appb-100004
    误差信号e(n)具体表达为：
    Figure PCTCN2021089536-appb-100005
    步骤(5)提取到的收敛后的自适应滤波器
    Figure PCTCN2021089536-appb-100006
    的N+1个系数，作为最终的射频指纹表达式为：
    Figure PCTCN2021089536-appb-100007
    其中，N表示N阶滤波器的阶数，n表示离散时间，k表示N阶滤波器的系数索引。
  3. 根据权利要求1所述的一种基于自适应滤波器系数的射频指纹统一表达方法，其特征在于：所述步骤(1)中的射频信号为无线信号，且接收机采样率f s大于奈奎斯特采样率，即要求过采样。
  4. 根据权利要求1所述的一种基于自适应滤波器系数的射频指纹统一表达方法，其特征在于，所述步骤(2)中的载波频偏和相偏估计方法包括但不限于前导复相关法或差分法。
  5. 根据权利要求1所述的一种基于自适应滤波器系数的射频指纹统一表达方法，其特征在于，所述步骤(3)中对包含射频指纹的基带信号y(n)进行符号解调和重新调制，是对完整的一帧信号进行处理，包括前导序列和随机数据序列。
  6. 根据权利要求1所述的一种基于自适应滤波器系数的射频指纹统一表达方法，其特征在于，所述步骤(4)中，对RFF(z)的系数进行调整时采用的自适应算法为迫零算法、最速下降法、LMS算法、RLS算法中的任一种。
  7. 根据权利要求6所述的一种基于自适应滤波器系数的射频指纹统一表达方法，其特征在于，所述步骤(4)中，使用LMS算法对RFF(z)的系数进行调整时，使得e(n)的均方误差最小，得到自适应滤波器的最终系数为：
    Figure PCTCN2021089536-appb-100008
    Figure PCTCN2021089536-appb-100009
    是使e(n)e *(n)具有最小值的RFF(z)的值的集合，e *(n)表示误差信号e(n)的复共轭。
  8. 一种电子设备，其特征在于，包括：
    信号接收模块，用于接收射频信号并进行采样和下变频处理，获得基带信号；
    第一处理模块，用于对基带信号进行信号同步、载波频偏和相偏估计，得到包含射频指纹的基带信号；
    第二处理模块，用于对包含射频指纹的基带信号进行符号解调和重新调制，得到不含射频指纹的理想基带信号；
    自适应计算模块，将射频指纹统一表达为z域的N阶滤波器，采用自适应算法对用于表达射频指纹的N阶滤波器的系数进行调整，调整结果收敛使得N阶滤波器的输出信号和包含射频指纹的基带信号的误差信号最小，N阶滤波器的 输入信号为理想基带信号；
    射频指纹提取模块，用于提取出收敛后的自适应滤波器的N+1个系数作为最终的射频指纹。
  9. 根据权利要求8所述的一种电子设备，其特征在于：所述自适应计算模块中，将射频指纹统一表达为z域的N阶滤波器RFF(z)，第N阶滤波器RFF(z)的表达式为：
    Figure PCTCN2021089536-appb-100010
    误差信号e(n)具体表达为：
    Figure PCTCN2021089536-appb-100011
    射频指纹提取模块中，最终的射频指纹表达式为：
    Figure PCTCN2021089536-appb-100012
    其中，N表示N阶滤波器的阶数，n表示离散时间，k表示N阶滤波器的系数索引。
  10. 根据权利要求8所述的一种电子设备，其特征在于：所述信号接收模块中，接收的射频信号为无线信号，第一处理模块使用本地的前导符信号和接收的基带信号信号进行复相关和差分处理，实现同步、载波频偏和相偏估计。
PCT/CN2021/089536 2021-04-16 2021-04-25 基于自适应滤波器系数的射频指纹统一表达方法及电子设备 WO2022217641A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110408422.2 2021-04-16
CN202110408422.2A CN113437949B (zh) 2021-04-16 2021-04-16 基于自适应滤波器系数的射频指纹统一表达方法及电子设备

Publications (1)

Publication Number Publication Date
WO2022217641A1 true WO2022217641A1 (zh) 2022-10-20

Family

ID=77753322

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/089536 WO2022217641A1 (zh) 2021-04-16 2021-04-25 基于自适应滤波器系数的射频指纹统一表达方法及电子设备

Country Status (2)

Country Link
CN (1) CN113437949B (zh)
WO (1) WO2022217641A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115396264B (zh) * 2022-08-22 2024-02-09 网络通信与安全紫金山实验室 射频指纹提取方法、装置和计算机设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102693411A (zh) * 2011-03-25 2012-09-26 南通大学 基于射频指纹的无线发射机的识别方法
CN103051458A (zh) * 2012-12-27 2013-04-17 南通大学 一种从bpsk信号中获取稳健射频指纹的方法
CN109921886A (zh) * 2019-01-28 2019-06-21 东南大学 一种鲁棒的低功耗设备射频指纹识别方法
CN111884974A (zh) * 2020-07-16 2020-11-03 东南大学 一种基于频谱分解的宽带信号射频指纹特征提取方法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7421376B1 (en) * 2001-04-24 2008-09-02 Auditude, Inc. Comparison of data signals using characteristic electronic thumbprints

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102693411A (zh) * 2011-03-25 2012-09-26 南通大学 基于射频指纹的无线发射机的识别方法
CN103051458A (zh) * 2012-12-27 2013-04-17 南通大学 一种从bpsk信号中获取稳健射频指纹的方法
CN109921886A (zh) * 2019-01-28 2019-06-21 东南大学 一种鲁棒的低功耗设备射频指纹识别方法
CN111884974A (zh) * 2020-07-16 2020-11-03 东南大学 一种基于频谱分解的宽带信号射频指纹特征提取方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
LI YUSHAN; CHEN LIUFEI; CHEN JIE; XIE FEIYI; CHEN SONGLIN; WEN HONG: "A Low Complexity Feature Extraction for the RF Fingerprinting Process", 2018 IEEE CONFERENCE ON COMMUNICATIONS AND NETWORK SECURITY (CNS), 30 May 2018 (2018-05-30), pages 1 - 2, XP033383958, DOI: 10.1109/CNS.2018.8433156 *
LIU YANPING, JINPENG TIAN, YONG CHEN: "RF Fingerprinting Identification with Feature Fusion", JOURNAL OF SHANGHAI UNIVERSITY (NATURAL SCIENCE EDITION), vol. 23, no. 3, 30 July 2017 (2017-07-30), pages 408 - 413, XP055977287, ISSN: 1007-2861, DOI: 10.12066/j.issn.1007-2861.1624 *
WILSON AARON J.; REISING DONALD R.; LOVELESS T. DANIEL: "Integration of Matched Filtering within the RF-DNA Fingerprinting Process", 2019 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM), 9 December 2019 (2019-12-09), pages 1 - 6, XP033722922, DOI: 10.1109/GLOBECOM38437.2019.9014225 *

Also Published As

Publication number Publication date
CN113437949B (zh) 2022-05-13
CN113437949A (zh) 2021-09-24

Similar Documents

Publication Publication Date Title
Anttila et al. Circularity-based I/Q imbalance compensation in wideband direct-conversion receivers
US5848105A (en) GMSK signal processors for improved communications capacity and quality
KR100668125B1 (ko) 적응성 사전-등화 방법 및 장치
US7031405B1 (en) Carrier phase estimation based on single-axis constant modulus cost criterion and Bussgang criteria
US7200180B2 (en) Data transceiver with filtering and precoding
JP4672229B2 (ja) デジタル通信経路上において適応型チャネル等化のトレーニング段階を実施するための方法および装置
US6205170B1 (en) Transmission/reception unit with bidirectional equalization
JPS61501954A (ja) ディジタル無線チャンネルのための相互ポ−ラリゼ−ション干渉除去法
WO2007088481A2 (en) Digital communications receiver
JP2000232399A (ja) 干渉及び非線形歪みをフィルタリングするための方法及び装置
US6088389A (en) System and method for training a plurality of equalizers and a modem employing the system or method
US20190245569A1 (en) Method for compensation of phase noise effect on data transmission in radio channel
EP3850805A1 (en) Radio frequency impairments compensator for broadband quadrature-conversion architectures
US7305048B2 (en) Burst mode receiver and method for stable reception of packet data on telephone line
WO2022217641A1 (zh) 基于自适应滤波器系数的射频指纹统一表达方法及电子设备
WO2019106447A1 (en) Ethernet transceiver with phy-level signal-loss detector
CN108768905B (zh) 一种基于混沌成型和混沌匹配滤波器的无线通信方法
US6879639B1 (en) Data transceiver with filtering and precoding
EP1219022B1 (en) Apparatus and method for substantially eliminating a near-channel interfering amplitude modulated signal
US7349469B1 (en) DC offset correction for constant modulus equalization
EP0956650A1 (en) Signal processing apparatus employing the spectral property of the signal
Shi Blind equalization and characteristic function based robust modulation recognition
JP6603252B2 (ja) 無線通信装置及び受信処理方法
Phukan et al. Blind Equalization for Classification of Digital Modulations
JP4511536B2 (ja) 適応的直交性訂正のための装置及び方法

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: 21936515

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21936515

Country of ref document: EP

Kind code of ref document: A1