CN115441970A - A Broadband Signal Detection Method Based on Scale Iteration and Spectrum Compensation - Google Patents

Abstract

Translated fromChinese

本发明属于无线通信技术领域，尤其涉及一种基于尺度迭代和频谱补偿的宽带信号检测方法，针对现有的宽带信号的检测算法易受噪声影响、在低信噪比下检测效果较差、复杂度较高，检测时间长、需要大量先验信息等问题，现提出如下方案，包括以下步骤：步骤A:收信号经过AD采样得到宽带采样信号，经过数字信道化将宽带信号划分为合适带宽的多段信号；本发明的目的是针对宽带信号，利用形态学滤波带宽筛选的特性，使噪声基底估计更准确，避免不平坦噪底造成误检；利用平滑迭代克服信号中强脉冲和白噪声对信号参数估计造成的误差，通过频谱补偿改善频带混叠信号的检测性能，实现宽带信号的盲检测，提高检测准确度。

The invention belongs to the technical field of wireless communication, and in particular relates to a wideband signal detection method based on scale iteration and spectrum compensation. The existing wideband signal detection algorithm is easily affected by noise, and the detection effect is poor and complicated under low signal-to-noise ratio. High accuracy, long detection time, need a large amount of prior information, etc., the following scheme is proposed, including the following steps: Step A: The received signal is subjected to AD sampling to obtain a broadband sampling signal, and the broadband signal is divided into suitable bandwidth through digital channelization Multi-segment signals; the purpose of the present invention is to use the characteristics of morphological filtering bandwidth screening for broadband signals to make the noise floor estimate more accurate, avoiding false detection caused by uneven noise floors; using smooth iteration to overcome the impact of strong pulses and white noise on the signal The error caused by parameter estimation is improved by spectral compensation to improve the detection performance of frequency band aliasing signals, realize blind detection of broadband signals, and improve detection accuracy.

Description

Translated fromChinese

一种基于尺度迭代和频谱补偿的宽带信号检测方法A Broadband Signal Detection Method Based on Scale Iteration and Spectrum Compensation

技术领域technical field

本发明属于无线通信技术领域，尤其涉及一种基于尺度迭代和频谱补偿的宽带信号检测方法。The invention belongs to the technical field of wireless communication, and in particular relates to a wideband signal detection method based on scale iteration and spectrum compensation.

背景技术Background technique

宽带信号检测的目的是在宽带频谱上快速找到信号，并估计出各个信号中心频率和带宽，为后续信号分析与处理提供先验信息。在各类民用和军用通信系统中均有大量应用，例如卫星频谱检测、雷达信号扫描以及战场信息的感知等均需要对宽带频谱进行检测。The purpose of wideband signal detection is to quickly find signals on the wideband spectrum, estimate the center frequency and bandwidth of each signal, and provide prior information for subsequent signal analysis and processing. There are a large number of applications in various civil and military communication systems, such as satellite spectrum detection, radar signal scanning, and battlefield information perception, all of which require detection of broadband spectrum.

然而，在非合作通信中，通常要在先验信息缺失、电磁环境复杂以及存在不平坦有色噪声基底等条件下完成频谱检测，这就要求宽带信号检测算法具备在低信噪比下无需先验信息实现盲检测并具备抗不平坦噪底的功能。However, in non-cooperative communication, it is usually necessary to complete spectrum detection under the conditions of lack of prior information, complex electromagnetic environment, and uneven colored noise floor. The information realizes blind detection and has the function of resisting uneven noise floor.

针对宽带信号的检测算法通常有如下几种：一是信号能量检测法，如Urkowitz(1967)提出未知确定性信号的能量检测方法、黎严(2017)提出改进的多天线信号能量检测方案等，这类方法具有抗频偏的优点，但易受噪声影响，在低信噪比下检测效果较差；二是循环平稳检测法，如A.Tani(2016)提出的基于二阶循环平稳特性算法，优点是检测性能较好，但复杂度较高，检测时间长；三是匹配滤波器检测法，如Z.Zhang(2010)提出的提出了平行匹配滤波器与分段滤波器相结合的检测方法，这类方法在理论上是一种最优的信号检测方法，但是需要载频、带宽、调制方式等大量先验信息，在实际中很难实现，不适用于非协作通信。另外，近年来还提出了一些新型算法，信号检测性能较经典算法有了很大的提升。张洋(2016)提出的梯度双阈值算法对宽带功率谱分段计算梯度,再进行自适应双阈值检测。齐佩汉(2014)提出的功率谱对消法利用分段频带内部分谱线强度和与全部谱线强度和的比值作为检验统计量进行信号存在性的判断。There are usually several detection algorithms for broadband signals: one is the signal energy detection method, such as Urkowitz (1967) proposed an energy detection method for unknown deterministic signals, Li Yan (2017) proposed an improved multi-antenna signal energy detection scheme, etc. This type of method has the advantage of anti-frequency offset, but it is easily affected by noise, and the detection effect is poor under low signal-to-noise ratio; the second is the cyclostationary detection method, such as the algorithm based on the second-order cyclostationary characteristic proposed by A. Tani (2016) , the advantage is that the detection performance is better, but the complexity is high and the detection time is long; the third is the matched filter detection method, such as Z. Zhang (2010) proposed a detection method combining parallel matched filters and segmented filters This type of method is an optimal signal detection method in theory, but it requires a lot of prior information such as carrier frequency, bandwidth, modulation mode, etc., which is difficult to implement in practice and is not suitable for non-cooperative communication. In addition, some new algorithms have been proposed in recent years, and the signal detection performance has been greatly improved compared with the classical algorithms. The gradient double-threshold algorithm proposed by Zhang Yang (2016) calculates the gradient of the broadband power spectrum segmentally, and then performs adaptive double-threshold detection. The power spectrum cancellation method proposed by Qi Peihan (2014) uses the ratio of the partial spectral line intensity sum to the entire spectral line intensity sum in the segmented frequency band as the test statistic to judge the existence of the signal.

发明内容Contents of the invention

针对现有的宽带信号的检测算法易受噪声影响、在低信噪比下检测效果较差、复杂度较高，检测时间长、需要大量先验信息等问题，而提出了一种基于尺度迭代和频谱补偿的宽带信号检测方法。Aiming at the problems that existing broadband signal detection algorithms are susceptible to noise, poor detection effect at low signal-to-noise ratio, high complexity, long detection time, and need a large amount of prior information, a scale-iterative algorithm based on scale iteration is proposed. and spectrally compensated broadband signal detection method.

为了实现上述目的，本发明采用了如下技术方案：In order to achieve the above object, the present invention adopts the following technical solutions:

一种基于尺度迭代和频谱补偿的宽带信号检测方法，包括如下步骤：A wideband signal detection method based on scale iteration and spectrum compensation, comprising the following steps:

步骤A:接收信号经过AD采样得到宽带采样信号，经过数字信道化将宽带信号划分为合适带宽的多段信号；Step A: the received signal obtains a broadband sampling signal through AD sampling, and divides the broadband signal into multi-section signals of suitable bandwidth through digital channelization;

步骤B:取数段宽带信号，据各段信号功率谱密度计算最大功率谱；Step B: take several sections of broadband signals, and calculate the maximum power spectrum according to the power spectral density of each section of signals;

步骤C:利用尺度迭代形态学滤波算法估计并去除有色噪声基底，并计算出白噪声功率门限；Step C: Utilize the scale iterative morphological filtering algorithm to estimate and remove the colored noise base, and calculate the white noise power threshold;

步骤D:检测功率谱中幅度最高的信号，估计出其中心频率、带宽以及信号频带起止位置，并根据补偿算法将信号剔除并完成补偿每次检出一个信号后，根据两次信号检测前后残余信号能量差值高于预设能量门限，则以补偿后的功率谱再次进行步骤D，否则信号检测结束。Step D: Detect the signal with the highest amplitude in the power spectrum, estimate its center frequency, bandwidth, and the start and end positions of the signal frequency band, and eliminate the signal according to the compensation algorithm and complete the compensation. After each signal is detected, the residual before and after two signal detections is used If the signal energy difference is higher than the preset energy threshold, step D is performed again with the compensated power spectrum, otherwise, the signal detection ends.

优选的，利用形态学滤波带宽筛选的特性，对功率谱进行形态学滤波迭代，使噪声基底估计更准确，利用平滑迭代克服信号中强脉冲和白噪声对信号参数估计造成的误差，通过频谱补偿改善频带混叠信号的检测性能。Preferably, using the characteristics of morphological filtering bandwidth screening, morphological filtering iterations are performed on the power spectrum to make the noise floor estimation more accurate, and smooth iterations are used to overcome the errors caused by strong pulses and white noise in the signal to signal parameter estimation, and through spectrum compensation Improved detection performance for band aliased signals.

优选的，所述步骤A具体包括如下步骤：Preferably, said step A specifically includes the following steps:

A1：计算实信号信道化子信道的中心频率

D为数据抽取倍数，整个频带被分为D个实信号的对称子带；A1: Calculate the center frequency of the channelized sub-channel of the real signal

D is the data extraction multiple, and the entire frequency band is divided into D symmetrical sub-bands of real signals;

A2：抗混叠滤波器h_LP(n)为低通FIR滤波器,K为信道个数,D为抽取因子,且N与D具有整数倍关系,即K＝FD,则传统信道化结构中的第k个子信道的输出为：A2: Anti-aliasing filter h_LP (n) is a low-pass FIR filter, K is the number of channels, D is the decimation factor, and N and D have an integer multiple relationship, that is, K=FD, then in the traditional channelization structure The output of the kth subchannel of is:

A3：将上式改写为多相滤波结构，得到第k个子信道的输出表达式为：A3: Rewrite the above formula into a polyphase filter structure, and the output expression of the kth sub-channel is obtained as:

A4：将中心频率ω_k代入式中,此时第k个子信道的输出表达式为：A4: Substitute the center frequency ω_k into the formula, and the output expression of the kth sub-channel at this time is:

优选的，所述步骤B中，将接收数据进行m次分段，每段长度为L，段间间隔为K，提取数据可以矩阵形式表示：Preferably, in the step B, the received data is segmented m times, the length of each segment is L, and the interval between segments is K, and the extracted data can be expressed in matrix form:

对每段数据以周期图法估计出各段数据的功率谱P_i(f)，即For each piece of data, the power spectrum P_i (f) of each piece of data is estimated by the periodogram method, namely

其中，w(n)是所选窗函数,求出各分段信号功率谱后在每个频率点处求出功率谱的最大值，即Among them, w(n) is the selected window function. After calculating the power spectrum of each segment signal, the maximum value of the power spectrum is calculated at each frequency point, that is

优选的，所述步骤C具体包括如下步骤：Preferably, said step C specifically includes the following steps:

B1:尺度B_bi取值最小检测带宽B_min为进行一次形态学滤波PN₁(f)；B1: The minimum detection bandwidth B_min of the scale B_bi is to perform a morphological filter PN₁ (f);

B2:尺度B_bi取值上一次滤波尺度B_bi-1的二倍进行一次形态学滤波，若尺度大于最大检测带宽B_max，则以B_max为尺度进行一次形态学滤波，第i次滤波结果为PN_i(f)；B2: The value of the scale B_bi is twice the value of the previous filtering scale B_bi-1 to perform a morphological filter. If the scale is greater than the maximum detection bandwidth B_max , perform a morphological filter with B_max as the scale. The i-th filtering result is PN_i (f);

B3:将第i次和第i-1次相邻两次滤波结果差分，差分结果为ΔPN_i，逐频点比较两次差分结果，若ΔPN_i(f)＞ΔPN_i-1(f)，则将频点f处尺度更新为B_bi，否则尺度保持不变；B3: Differentiate the i-th and i-1th adjacent two filtering results, the difference result is ΔPN_i , compare the two difference results frequency by point, if ΔPN_i (f) > ΔPN_i-1 (f), Then update the scale at the frequency point f to B_bi , otherwise the scale remains unchanged;

B4:若B_bi＝B_max，则停止迭代；B4: if B_bi =B_max , then stop iteration;

B5:各个频点f处以迭代确定的尺度B_bi(f)进行一次形态学滤波估计出有色噪声基底；B5: Carry out a morphological filter with the iteratively determined scale B_bi (f) at each frequency point f to estimate the colored noise floor;

B6:以原功率谱减B5估计出噪声基底得到噪底平坦的功率谱；B6: Estimate the noise floor with the original power spectrum minus B5 to obtain a flat power spectrum with a noise floor;

B7:将功率谱转换为对数功率谱，在功率谱上由0至最大值划定适当个数的功率带并统计落在各带内的频点数，功率带个数越多，白噪声门限估计越准确，相邻功率带频点数分别差分，以差分结果最大的功率带的上限作为白噪声功率门限。B7: Convert the power spectrum to a logarithmic power spectrum, delimit an appropriate number of power bands from 0 to the maximum value on the power spectrum and count the number of frequency points falling in each band, the more the number of power bands, the white noise threshold The more accurate the estimation, the frequency points of adjacent power bands are differentiated, and the upper limit of the power band with the largest difference result is used as the white noise power threshold.

优选的，所述步骤D中具体包括如下步骤：Preferably, the step D specifically includes the following steps:

C1:在功率谱中寻找最大值位置f_max作为中心频率f_Ctr，在中心频率幅度P(f_Ctr)-3dB处计算出3dB带宽B_3dB，带宽左边界为f_3dBL，右边界为f_3dBR；C1: Find the maximum position f_max in the power spectrum as the center frequency f_Ctr , calculate the 3dB bandwidth B_3dB at the center frequency amplitude P(f_Ctr )-3dB, the left boundary of the bandwidth is f_3dBL , and the right boundary is f_3dBR ;

C2:将功率谱在[f_3dBL,f_3dBR]以尺度λ_sm＝0.1*B_3dB进行平滑滤波处理后返回C1，循环五次后进入C3；C2: return the power spectrum to C1 after smoothing and filtering the power spectrum at [f_3dBL , f_3dBR ] with the scale λ_sm =0.1*B_3dB , and enter C3 after five cycles;

C3:经过五次循环平滑后，所得带宽即为3dB带宽，在最终的[f_3dBL,f_3dBR]范围内根据下式计算中心频率；C3: After five cycles of smoothing, the resulting bandwidth is the 3dB bandwidth, and the center frequency is calculated according to the following formula within the final [f_3dBL , f_3dBR ] range;

C4:在f_3dBL以左，f_3dBR以右寻找功率谱首个极小值点或零点作为信号主瓣边界，左边界为f_LS，右边界为f_RS，在[f_LS,f_RS]范围内将信号归零剔除，根据下式计算带外斜率k_LS、k_RS：C4: To the left of f_3dBL and to the right of f_3dBR , look for the first minimum point or zero point of the power spectrum as the signal main lobe boundary, the left boundary is f_LS , the right boundary is f_RS , in the range of [f_LS , f_RS ] The signal is returned to zero and eliminated, and the out-of-band slopes k_LS and k_RS are calculated according to the following formula:

C5:在f_LS，f_RS处分别以斜率k_LS、k_RS向已估计信号带内[f_LS,f_RS]延伸，至两线相交为止，以两条延长线代替已估计信号的带内谱线，减小对为未估计信号的损伤；C5: At f_LS and f_RS , extend to the estimated signal band [f_LS , f_RS ] with slopes k_LS and k_RS respectively until the two lines intersect, and replace the estimated signal band with two extension lines Spectral lines, reducing damage to unestimated signals;

C6:若C4已估计能量低于最小判定能量或带宽小于最小检测带宽，则将其舍弃，若残余功率谱最大幅度不高于白噪声门限，则信号检测完成。C6: If the energy estimated by C4 is lower than the minimum decision energy or the bandwidth is smaller than the minimum detection bandwidth, it is discarded. If the maximum amplitude of the residual power spectrum is not higher than the white noise threshold, the signal detection is completed.

本发明所具有的有益效果为：针对宽带信号，利用形态学滤波带宽筛选的特性，对功率谱进行形态学滤波迭代，使噪声基底估计更准确，避免不平坦噪底造成误检；利用平滑迭代克服信号中强脉冲和白噪声对信号参数估计造成的误差，通过频谱补偿改善频带混叠信号的检测性能，实现宽带信号的盲检测，提高检测准确度。The beneficial effects of the present invention are as follows: for broadband signals, using the characteristics of morphological filtering bandwidth screening, morphological filtering iteration is performed on the power spectrum, so that the noise floor estimation is more accurate, and false detection caused by uneven noise floor is avoided; Overcome the error caused by the strong pulse and white noise in the signal to the signal parameter estimation, improve the detection performance of the frequency band aliasing signal through spectrum compensation, realize the blind detection of the broadband signal, and improve the detection accuracy.

附图说明Description of drawings

图1为一种基于尺度迭代和频谱补偿的宽带信号检测方法的步骤流程图；Fig. 1 is a flow chart of the steps of a broadband signal detection method based on scale iteration and spectrum compensation;

图2为数字信道化结构；Fig. 2 is a digital channelization structure;

图3为宽带信号频谱示意图；Fig. 3 is the schematic diagram of broadband signal frequency spectrum;

图4为信号检测环节流程；Figure 4 is the process flow of the signal detection link;

图5为有色噪声基底估计结果示意图；Fig. 5 is a schematic diagram of the estimation result of the colored noise floor;

图6为信号检测结果示意图；Fig. 6 is a schematic diagram of signal detection results;

图7为在不同信噪比下的信号正确概率。Fig. 7 shows the signal correct probability under different signal-to-noise ratios.

具体实施方式detailed description

下面将对本发明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。The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some of the embodiments of the present invention, but not all of them.

实施例一Embodiment one

如图1-7所示，种基于尺度迭代和频谱补偿的宽带信号检测方法，包括如下步骤：As shown in Figure 1-7, a wideband signal detection method based on scale iteration and spectrum compensation includes the following steps:

步骤A:接收信号经过AD采样得到宽带采样信号，经过数字信道化将宽带信号划分为合适带宽的多段信号；Step A: the received signal obtains a broadband sampling signal through AD sampling, and divides the broadband signal into multi-section signals of suitable bandwidth through digital channelization;

步骤B:取数段宽带信号，据各段信号功率谱密度计算最大功率谱；Step B: take several sections of broadband signals, and calculate the maximum power spectrum according to the power spectral density of each section of signals;

步骤C:利用尺度迭代形态学滤波算法估计并去除有色噪声基底，并计算出白噪声功率门限；Step C: Utilize the scale iterative morphological filtering algorithm to estimate and remove the colored noise base, and calculate the white noise power threshold;

步骤D:检测功率谱中幅度最高的信号，估计出其中心频率、带宽以及信号频带起止位置，并根据补偿算法将信号剔除并完成补偿每次检出一个信号后，根据两次信号检测前后残余信号能量差值高于预设能量门限，则以补偿后的功率谱再次进行步骤D，否则信号检测结束。Step D: Detect the signal with the highest amplitude in the power spectrum, estimate its center frequency, bandwidth, and the start and end positions of the signal frequency band, and eliminate the signal according to the compensation algorithm and complete the compensation. After each signal is detected, the residual before and after two signal detections is used If the signal energy difference is higher than the preset energy threshold, step D is performed again with the compensated power spectrum, otherwise, the signal detection ends.

本实施例中，所述步骤A具体包括如下步骤：In this embodiment, the step A specifically includes the following steps:

A1：计算实信号信道化子信道的中心频率

D为数据抽取倍数，整个频带被分为D个实信号的对称子带；A1: Calculate the center frequency of the channelized sub-channel of the real signal

D is the data extraction multiple, and the entire frequency band is divided into D symmetrical sub-bands of real signals;

A2：抗混叠滤波器h_LP(n)为低通FIR滤波器,K为信道个数,D为抽取因子,且N与D具有整数倍关系,即K＝FD,则传统信道化结构中的第k个子信道的输出为：A2: Anti-aliasing filter h_LP (n) is a low-pass FIR filter, K is the number of channels, D is the decimation factor, and N and D have an integer multiple relationship, that is, K=FD, then in the traditional channelization structure The output of the kth subchannel of is:

A3：将上式改写为多相滤波结构，得到第k个子信道的输出表达式为：A3: Rewrite the above formula into a polyphase filter structure, and the output expression of the kth sub-channel is obtained as:

A4：将中心频率ω_k代入式中,此时第k个子信道的输出表达式为：A4: Substitute the center frequency ω_k into the formula, and the output expression of the kth sub-channel at this time is:

本实施例中，所述步骤B中，将接收数据进行m次分段，每段长度为L，段间间隔为K，提取数据可以矩阵形式表示：In this embodiment, in the step B, the received data is segmented m times, each segment length is L, and the interval between segments is K, and the extracted data can be expressed in matrix form:

对每段数据以周期图法估计出各段数据的功率谱P_i(f)，即For each piece of data, the power spectrum P_i (f) of each piece of data is estimated by the periodogram method, namely

其中，w(n)是所选窗函数，求出各分段信号功率谱后在每个频率点处求出功率谱的最大值，即Among them, w(n) is the selected window function, after obtaining the power spectrum of each segmented signal, the maximum value of the power spectrum is obtained at each frequency point, that is

本实施例中，所述步骤C具体包括如下步骤：In this embodiment, the step C specifically includes the following steps:

B1:尺度B_bi取值最小检测带宽B_min为进行一次形态学滤波PN₁(f)；B1: The minimum detection bandwidth B_min of the scale B_bi is to perform a morphological filter PN₁ (f);

B2:尺度B_bi取值上一次滤波尺度B_bi-1的二倍进行一次形态学滤波，若尺度大于最大检测带宽B_max，则以B_max为尺度进行一次形态学滤波，第i次滤波结果为PN_i(f)；B2: The value of the scale B_bi is twice the value of the previous filtering scale B_bi-1 to perform a morphological filter. If the scale is greater than the maximum detection bandwidth B_max , perform a morphological filter with B_max as the scale. The i-th filtering result is PN_i (f);

B3:将第i次和第i-1次相邻两次滤波结果差分，差分结果为ΔPN_i，逐频点比较两次差分结果，若ΔPN_i(f)＞ΔPN_i-1(f)，则将频点f处尺度更新为B_bi，否则尺度保持不变；B3: Differentiate the i-th and i-1th adjacent two filtering results, the difference result is ΔPN_i , compare the two difference results frequency by point, if ΔPN_i (f) > ΔPN_i-1 (f), Then update the scale at the frequency point f to B_bi , otherwise the scale remains unchanged;

B4:若B_bi＝B_max，则停止迭代；B4: if B_bi =B_max , then stop iteration;

B5:各个频点f处以迭代确定的尺度B_bi(f)进行一次形态学滤波估计出有色噪声基底；B5: Carry out a morphological filter with the iteratively determined scale B_bi (f) at each frequency point f to estimate the colored noise floor;

B6:以原功率谱减B5估计出噪声基底得到噪底平坦的功率谱；B6: Estimate the noise floor with the original power spectrum minus B5 to obtain a flat power spectrum with a noise floor;

B7:将功率谱转换为对数功率谱，在功率谱上由0至最大值划定适当个数的功率带并统计落在各带内的频点数，功率带个数越多，白噪声门限估计越准确，相邻功率带频点数分别差分，以差分结果最大的功率带的上限作为白噪声功率门限。B7: Convert the power spectrum to a logarithmic power spectrum, delimit an appropriate number of power bands from 0 to the maximum value on the power spectrum and count the number of frequency points falling in each band, the more the number of power bands, the white noise threshold The more accurate the estimation, the frequency points of adjacent power bands are differentiated, and the upper limit of the power band with the largest difference result is used as the white noise power threshold.

本实施例中，所述步骤D中具体包括如下步骤：In this embodiment, the step D specifically includes the following steps:

C1:在功率谱中寻找最大值位置f_max作为中心频率f_Ctr，在中心频率幅度P(f_Ctr)-3dB处计算出3dB带宽B_3dB，带宽左边界为f_3dBL，右边界为f_3dBR；C1: Find the maximum position f_max in the power spectrum as the center frequency f_Ctr , calculate the 3dB bandwidth B_3dB at the center frequency amplitude P(f_Ctr )-3dB, the left boundary of the bandwidth is f_3dBL , and the right boundary is f_3dBR ;

C2:将功率谱在[f_3dBL,f_3dBR]以尺度λ_sm＝0.1*B_3dB进行平滑滤波处理后返回C1，循环五次后进入C3；C2: return the power spectrum to C1 after smoothing and filtering the power spectrum at [f_3dBL , f_3dBR ] with the scale λ_sm =0.1*B_3dB , and enter C3 after five cycles;

C3:经过五次循环平滑后，所得带宽即为3dB带宽，在最终的[f_3dBL,f_3dBR]范围内根据下式计算中心频率；C3: After five cycles of smoothing, the resulting bandwidth is the 3dB bandwidth, and the center frequency is calculated according to the following formula within the final [f_3dBL , f_3dBR ] range;

C4:在f_3dBL以左，f_3dBR以右寻找功率谱首个极小值点或零点作为信号主瓣边界，左边界为f_LS，右边界为f_RS，在[f_LS,f_RS]范围内将信号归零剔除，根据下式计算带外斜率k_LS、k_RS：C4: To the left of f_3dBL and to the right of f_3dBR , look for the first minimum point or zero point of the power spectrum as the signal main lobe boundary, the left boundary is f_LS , the right boundary is f_RS , in the range of [f_LS , f_RS ] The signal is returned to zero and eliminated, and the out-of-band slopes k_LS and k_RS are calculated according to the following formula:

C5:在f_LS，f_RS处分别以斜率k_LS、k_RS向已估计信号带内[f_LS,f_RS]延伸，至两线相交为止，以两条延长线代替已估计信号的带内谱线，减小对为未估计信号的损伤；C5: At f_LS and f_RS , extend to the estimated signal band [f_LS , f_RS ] with slopes k_LS and k_RS respectively until the two lines intersect, and replace the estimated signal band with two extension lines Spectral lines, reducing damage to unestimated signals;

C6:若C4已估计能量低于最小判定能量或带宽小于最小检测带宽，则将其舍弃，若残余功率谱最大幅度不高于白噪声门限，则信号检测完成。C6: If the energy estimated by C4 is lower than the minimum decision energy or the bandwidth is smaller than the minimum detection bandwidth, it is discarded. If the maximum amplitude of the residual power spectrum is not higher than the white noise threshold, the signal detection is completed.

实施例二Embodiment two

如图1-7所示，一种基于尺度迭代和频谱补偿的宽带信号检测方法，包括如下步骤：As shown in Figure 1-7, a wideband signal detection method based on scale iteration and spectrum compensation includes the following steps:

步骤A:接收信号经过AD采样得到宽带采样信号，经过数字信道化将宽带信号划分为合适带宽的多段信号；Step A: the received signal obtains a broadband sampling signal through AD sampling, and divides the broadband signal into multi-section signals of suitable bandwidth through digital channelization;

步骤B:取数段宽带信号，据各段信号功率谱密度计算最大功率谱；Step B: take several sections of broadband signals, and calculate the maximum power spectrum according to the power spectral density of each section of signals;

步骤C:利用尺度迭代形态学滤波算法估计并去除有色噪声基底，并计算出白噪声功率门限；Step C: Utilize the scale iterative morphological filtering algorithm to estimate and remove the colored noise base, and calculate the white noise power threshold;

步骤D:检测功率谱中幅度最高的信号，并根据补偿算法将信号剔除并完成补偿每次检出一个信号后，根据两次信号检测前后残余信号能量差值高于预设能量门限，则以补偿后的功率谱再次进行步骤D，否则信号检测结束。Step D: Detect the signal with the highest amplitude in the power spectrum, and remove the signal according to the compensation algorithm and complete the compensation. After each signal is detected, according to the residual signal energy difference before and after the two signal detections is higher than the preset energy threshold, then use The compensated power spectrum is subjected to step D again, otherwise the signal detection ends.

本实施例中，所述步骤A具体包括如下步骤：In this embodiment, the step A specifically includes the following steps:

A1：计算实信号信道化子信道的中心频率

D为数据抽取倍数，整个频带被分为D个实信号的对称子带；A1: Calculate the center frequency of the channelized sub-channel of the real signal

D is the data extraction multiple, and the entire frequency band is divided into D symmetrical sub-bands of real signals;

A2：抗混叠滤波器h_LP(n)为低通FIR滤波器,K为信道个数,D为抽取因子,且N与D具有整数倍关系,即K＝FD,则传统信道化结构中的第k个子信道的输出为：A2: Anti-aliasing filter h_LP (n) is a low-pass FIR filter, K is the number of channels, D is the decimation factor, and N and D have an integer multiple relationship, that is, K=FD, then in the traditional channelization structure The output of the kth subchannel of is:

A3：将上式改写为多相滤波结构，得到第k个子信道的输出表达式为：A3: Rewrite the above formula into a polyphase filter structure, and the output expression of the kth sub-channel is obtained as:

A4：将中心频率ω_k代入式中,此时第k个子信道的输出表达式为：A4: Substitute the center frequency ω_k into the formula, and the output expression of the kth sub-channel at this time is:

本实施例中，所述步骤B中，将接收数据进行m次分段，每段长度为L，段间间隔为K，提取数据可以矩阵形式表示：In this embodiment, in the step B, the received data is segmented m times, each segment length is L, and the interval between segments is K, and the extracted data can be expressed in matrix form:

对每段数据以周期图法估计出各段数据的功率谱P_i(f)，即For each piece of data, the power spectrum P_i (f) of each piece of data is estimated by the periodogram method, namely

其中，w(n)是所选窗函数，求出各分段信号功率谱后在每个频率点处求出功率谱的最大值，即Among them, w(n) is the selected window function, after obtaining the power spectrum of each segmented signal, the maximum value of the power spectrum is obtained at each frequency point, that is

本实施例中，所述步骤C具体包括如下步骤：In this embodiment, the step C specifically includes the following steps:

B1:尺度B_bi取值最小检测带宽B_min为进行一次形态学滤波PN₁(f)；B1: The minimum detection bandwidth B_min of the scale B_bi is to perform a morphological filter PN₁ (f);

B2:尺度B_bi取值上一次滤波尺度B_bi-1的二倍进行一次形态学滤波，第i次滤波结果为PN_i(f)；B2: The scale B_bi is twice the value of the previous filtering scale B_bi-1 to perform a morphological filtering, and the result of the i-th filtering is PN_i (f);

B3:将第i次和第i-1次相邻两次滤波结果差分，差分结果为ΔPN_i，若ΔPN_i(f)＞ΔPN_i-1(f)，则将频点f处尺度更新为B_bi，否则尺度保持不变；B3: Difference between the i-th and i-1th adjacent two filtering results, the difference result is ΔPN_i , if ΔPN_i (f) > ΔPN_i-1 (f), update the scale at the frequency point f to B_bi , otherwise the scale remains unchanged;

B4:若B_bi＝B_max，则停止迭代；B4: if B_bi =B_max , then stop iteration;

B5:各个频点f处以迭代确定的尺度B_bi(f)进行一次形态学滤波估计出有色噪声基底；B5: Carry out a morphological filter with the iteratively determined scale B_bi (f) at each frequency point f to estimate the colored noise floor;

B6:以原功率谱减B5估计出噪声基底得到噪底平坦的功率谱；B6: Estimate the noise floor with the original power spectrum minus B5 to obtain a flat power spectrum with a noise floor;

B7:将功率谱转换为对数功率谱，在功率谱上由0至最大值划定适当个数的功率带并统计落在各带内的频点数，功率带个数越多，白噪声门限估计越准确，相邻功率带频点数分别差分，以差分结果最大的功率带的上限作为白噪声功率门限。B7: Convert the power spectrum to a logarithmic power spectrum, delimit an appropriate number of power bands from 0 to the maximum value on the power spectrum and count the number of frequency points falling in each band, the more the number of power bands, the white noise threshold The more accurate the estimation, the frequency points of adjacent power bands are differentiated, and the upper limit of the power band with the largest difference result is used as the white noise power threshold.

本实施例中，所述步骤D中具体包括如下步骤：In this embodiment, the step D specifically includes the following steps:

C1:在功率谱中寻找最大值位置f_max作为中心频率f_Ctr，在中心频率幅度P(f_Ctr)-3dB处计算出3dB带宽B_3dB，带宽左边界为f_3dBL，右边界为f_3dBR；C1: Find the maximum position f_max in the power spectrum as the center frequency f_Ctr , calculate the 3dB bandwidth B_3dB at the center frequency amplitude P(f_Ctr )-3dB, the left boundary of the bandwidth is f_3dBL , and the right boundary is f_3dBR ;

C2:将功率谱在[f_3dBL,f_3dBR]以尺度λ_sm＝0.1*B_3dB进行平滑滤波处理后返回C1，循环五次后进入C3；C2: return the power spectrum to C1 after smoothing and filtering the power spectrum at [f_3dBL , f_3dBR ] with the scale λ_sm =0.1*B_3dB , and enter C3 after five cycles;

C3:经过五次循环平滑后，所得带宽即为3dB带宽，在最终的[f_3dBL,f_3dBR]范围内根据下式计算中心频率；C3: After five cycles of smoothing, the resulting bandwidth is the 3dB bandwidth, and the center frequency is calculated according to the following formula within the final [f_3dBL , f_3dBR ] range;

C4:在f_3dBL以左，f_3dBR以右寻找功率谱首个极小值点或零点作为信号主瓣边界，在[f_LS,f_RS]范围内将信号归零剔除，根据下式计算带外斜率k_LS、k_RS：C4: To the left of f_3dBL and to the right of f_3dBR , look for the first minimum point or zero point of the power spectrum as the main lobe boundary of the signal, return the signal to zero within the range of [f_LS , f_RS ], and calculate the band according to the following formula Outer slope k_LS , k_RS :

C5:在f_LS，f_RS处分别以斜率k_LS、k_RS向已估计信号带内[f_LS,f_RS]延伸，至两线相交为止，减小对为未估计信号的损伤；C5: At f_LS and f_RS , extend to the estimated signal band [f_LS , f_RS ] with slopes k_LS and k_RS respectively until the two lines intersect, reducing the damage to the unestimated signal;

C6:若C4已估计能量低于最小判定能量或带宽小于最小检测带宽，则将其舍弃，若残余功率谱最大幅度不高于白噪声门限，则信号检测完成。C6: If the energy estimated by C4 is lower than the minimum decision energy or the bandwidth is smaller than the minimum detection bandwidth, it is discarded. If the maximum amplitude of the residual power spectrum is not higher than the white noise threshold, the signal detection is completed.

实施例三Embodiment three

如图1-7所示，一种基于尺度迭代和频谱补偿的宽带信号检测方法，包括如下步骤：As shown in Figure 1-7, a wideband signal detection method based on scale iteration and spectrum compensation includes the following steps:

步骤A:接收信号经过AD采样得到宽带采样信号，经过数字信道化将宽带信号划分为合适带宽的多段信号；Step A: the received signal obtains a broadband sampling signal through AD sampling, and divides the broadband signal into multi-section signals of suitable bandwidth through digital channelization;

步骤B:取数段宽带信号，据各段信号功率谱密度计算最大功率谱；Step B: take several sections of broadband signals, and calculate the maximum power spectrum according to the power spectral density of each section of signals;

步骤C:利用尺度迭代形态学滤波算法估计并去除有色噪声基底，并计算出白噪声功率门限；Step C: Utilize the scale iterative morphological filtering algorithm to estimate and remove the colored noise base, and calculate the white noise power threshold;

步骤D:检测功率谱中幅度最高的信号，根据两次信号检测前后残余信号能量差值高于预设能量门限，则以补偿后的功率谱再次进行步骤D，否则信号检测结束。Step D: Detect the signal with the highest amplitude in the power spectrum. According to the residual signal energy difference before and after the two signal detections is higher than the preset energy threshold, then perform step D again with the compensated power spectrum, otherwise the signal detection ends.

本实施例中，所述步骤A具体包括如下步骤：In this embodiment, the step A specifically includes the following steps:

A1：计算实信号信道化子信道的中心频率

D为数据抽取倍数，整个频带被分为D个实信号的对称子带；A1: Calculate the center frequency of the channelized sub-channel of the real signal

D is the data extraction multiple, and the entire frequency band is divided into D symmetrical sub-bands of real signals;

A2：抗混叠滤波器h_LP(n)为低通FIR滤波器,K为信道个数,D为抽取因子,且N与D具有整数倍关系,即K＝FD,则传统信道化结构中的第k个子信道的输出为：A2: Anti-aliasing filter h_LP (n) is a low-pass FIR filter, K is the number of channels, D is the decimation factor, and N and D have an integer multiple relationship, that is, K=FD, then in the traditional channelization structure The output of the kth subchannel of is:

A4：将中心频率ω_k代入式中,此时第k个子信道的输出表达式为：A4: Substitute the center frequency ω_k into the formula, and the output expression of the kth sub-channel at this time is:

本实施例中，所述步骤B中，将接收数据进行m次分段，每段长度为L，段间间隔为K，提取数据可以矩阵形式表示：In this embodiment, in the step B, the received data is segmented m times, each segment length is L, and the interval between segments is K, and the extracted data can be expressed in matrix form:

对每段数据以周期图法估计出各段数据的功率谱P_i(f)，即For each piece of data, the power spectrum P_i (f) of each piece of data is estimated by the periodogram method, namely

其中，w(n)是所选窗函数，求出各分段信号功率谱后在每个频率点处求出功率谱的最大值，即Among them, w(n) is the selected window function, after obtaining the power spectrum of each segmented signal, the maximum value of the power spectrum is obtained at each frequency point, that is

本实施例中，所述步骤C具体包括如下步骤：In this embodiment, the step C specifically includes the following steps:

B1:尺度B_bi取值最小检测带宽B_min为进行一次形态学滤波PN₁(f)；B1: The minimum detection bandwidth B_min of the scale B_bi is to perform a morphological filter PN₁ (f);

B2:尺度B_bi取值上一次滤波尺度B_bi-1的二倍进行一次形态学滤波，若尺度大于最大检测带宽B_max，则以B_max为尺度进行一次形态学滤波，第i次滤波结果为PN_i(f)；B2: The value of the scale B_bi is twice the value of the previous filtering scale B_bi-1 to perform a morphological filter. If the scale is greater than the maximum detection bandwidth B_max , perform a morphological filter with B_max as the scale. The i-th filtering result is PN_i (f);

B3:将第i次和第i-1次相邻两次滤波结果差分，差分结果为ΔPN_i，逐频点比较两次差分结果，若ΔPN_i(f)＞ΔPN_i-1(f)，则将频点f处尺度更新为B_bi，否则尺度保持不变；B3: Differentiate the i-th and i-1th adjacent two filtering results, the difference result is ΔPN_i , compare the two difference results frequency by point, if ΔPN_i (f) > ΔPN_i-1 (f), Then update the scale at the frequency point f to B_bi , otherwise the scale remains unchanged;

B4:各个频点f处以迭代确定的尺度B_bi(f)进行一次形态学滤波估计出有色噪声基底；B4: Carry out a morphological filter with the iteratively determined scale B_bi (f) at each frequency point f to estimate the colored noise floor;

B5:以原功率谱减B5估计出噪声基底得到噪底平坦的功率谱；B5: Estimate the noise floor with the original power spectrum minus B5 to obtain a flat power spectrum with a noise floor;

B6:将功率谱转换为对数功率谱，在功率谱上由0至最大值划定适当个数的功率带并统计落在各带内的频点数，以差分结果最大的功率带的上限作为白噪声功率门限。B6: Convert the power spectrum to a logarithmic power spectrum, delineate an appropriate number of power bands from 0 to the maximum value on the power spectrum and count the number of frequency points falling in each band, and take the upper limit of the power band with the largest difference result as White noise power threshold.

本实施例中，所述步骤D中具体包括如下步骤：In this embodiment, the step D specifically includes the following steps:

C1:在功率谱中寻找最大值位置f_max作为中心频率f_Ctr，在中心频率幅度P(f_Ctr)-3dB处计算出3dB带宽B_3dB，带宽左边界为f_3dBL，右边界为f_3dBR；C1: Find the maximum position f_max in the power spectrum as the center frequency f_Ctr , calculate the 3dB bandwidth B_3dB at the center frequency amplitude P(f_Ctr )-3dB, the left boundary of the bandwidth is f_3dBL , and the right boundary is f_3dBR ;

C2:将功率谱在[f_3dBL,f_3dBR]以尺度λ_sm＝0.1*B_3dB进行平滑滤波处理后返回C1，循环五次后进入C3；C2: return the power spectrum to C1 after smoothing and filtering the power spectrum at [f_3dBL , f_3dBR ] with the scale λ_sm =0.1*B_3dB , and enter C3 after five cycles;

C3:所得带宽即为3dB带宽，在最终的[f_3dBL,f_3dBR]范围内根据下式计算中心频率；C3: The resulting bandwidth is the 3dB bandwidth, and the center frequency is calculated according to the following formula within the final [f_3dBL , f_3dBR ] range;

C4:在f_3dBL以左，f_3dBR以右寻找功率谱首个极小值点或零点作为信号主瓣边界，左边界为f_LS，右边界为f_RS，在[f_LS,f_RS]范围内将信号归零剔除，根据下式计算带外斜率k_LS、k_RS：C4: To the left of f_3dBL and to the right of f_3dBR , look for the first minimum point or zero point of the power spectrum as the signal main lobe boundary, the left boundary is f_LS , the right boundary is f_RS , in the range of [f_LS , f_RS ] The signal is returned to zero and eliminated, and the out-of-band slopes k_LS and k_RS are calculated according to the following formula:

C5:在f_LS，f_RS处分别以斜率k_LS、k_RS向已估计信号带内[f_LS,f_RS]延伸，至两线相交为止，以两条延长线代替已估计信号的带内谱线，减小对为未估计信号的损伤；C5: At f_LS and f_RS , extend to the estimated signal band [f_LS , f_RS ] with slopes k_LS and k_RS respectively until the two lines intersect, and replace the estimated signal band with two extension lines Spectral lines, reducing damage to unestimated signals;

C6:若C4已估计能量低于最小判定能量或带宽小于最小检测带宽，则将其舍弃，若残余功率谱最大幅度不高于白噪声门限，则信号检测完成。C6: If the energy estimated by C4 is lower than the minimum decision energy or the bandwidth is smaller than the minimum detection bandwidth, it is discarded. If the maximum amplitude of the residual power spectrum is not higher than the white noise threshold, the signal detection is completed.

实施例四Embodiment four

如图1-7所示，一种基于尺度迭代和频谱补偿的宽带信号检测方法，包括如下步骤：As shown in Figure 1-7, a wideband signal detection method based on scale iteration and spectrum compensation includes the following steps:

步骤A:接收信号经过AD采样得到宽带采样信号；Step A: the received signal obtains a broadband sampling signal through AD sampling;

步骤B:取数段宽带信号，据各段信号功率谱密度计算最大功率谱；Step B: take several sections of broadband signals, and calculate the maximum power spectrum according to the power spectral density of each section of signals;

步骤C:利用尺度迭代形态学滤波算法估计并去除有色噪声基底，并计算出白噪声功率门限；Step C: Utilize the scale iterative morphological filtering algorithm to estimate and remove the colored noise base, and calculate the white noise power threshold;

步骤D:检测功率谱中幅度最高的信号，估计出其中心频率、带宽以及信号频带起止位置，并根据补偿算法将信号剔除并完成补偿每次检出一个信号后，根据两次信号检测前后残余信号能量差值高于预设能量门限，则以补偿后的功率谱再次进行步骤D，否则信号检测结束。Step D: Detect the signal with the highest amplitude in the power spectrum, estimate its center frequency, bandwidth, and the start and end positions of the signal frequency band, and eliminate the signal according to the compensation algorithm and complete the compensation. After each signal is detected, the residual before and after two signal detections is used If the signal energy difference is higher than the preset energy threshold, step D is performed again with the compensated power spectrum, otherwise, the signal detection ends.

本实施例中，所述步骤A具体包括如下步骤：In this embodiment, the step A specifically includes the following steps:

A1：计算实信号信道化子信道的中心频率

D为数据抽取倍数，整个频带被分为D个实信号的对称子带；A1: Calculate the center frequency of the channelized sub-channel of the real signal

D is the data extraction multiple, and the entire frequency band is divided into D symmetrical sub-bands of real signals;

A2：抗混叠滤波器h_LP(n)为低通FIR滤波器,K为信道个数,D为抽取因子,且N与D具有整数倍关系,即K＝FD,则传统信道化结构中的第k个子信道的输出为：A2: Anti-aliasing filter h_LP (n) is a low-pass FIR filter, K is the number of channels, D is the decimation factor, and N and D have an integer multiple relationship, that is, K=FD, then in the traditional channelization structure The output of the kth subchannel of is:

A3：将上式改写为多相滤波结构，得到第k个子信道的输出表达式为：A3: Rewrite the above formula into a polyphase filter structure, and the output expression of the kth sub-channel is obtained as:

A4：将中心频率ω_k代入式中,此时第k个子信道的输出表达式为：A4: Substitute the center frequency ω_k into the formula, and the output expression of the kth sub-channel at this time is:

本实施例中，所述步骤B中，将接收数据进行m次分段，每段长度为L，段间间隔为K，提取数据可以矩阵形式表示：In this embodiment, in the step B, the received data is segmented m times, each segment length is L, and the interval between segments is K, and the extracted data can be expressed in matrix form:

对每段数据以周期图法估计出各段数据的功率谱P_i(f)，即For each piece of data, the power spectrum P_i (f) of each piece of data is estimated by the periodogram method, namely

其中，w(n)是所选窗函数，求出各分段信号功率谱后在每个频率点处求出功率谱的最大值，即Among them, w(n) is the selected window function, after obtaining the power spectrum of each segmented signal, the maximum value of the power spectrum is obtained at each frequency point, that is

本实施例中，所述步骤C具体包括如下步骤：In this embodiment, the step C specifically includes the following steps:

B1:尺度B_bi取值最小检测带宽B_min为进行一次形态学滤波PN₁(f)；B1: The minimum detection bandwidth B_min of the scale B_bi is to perform a morphological filter PN₁ (f);

B2:尺度B_bi取值上一次滤波尺度B_bi-1的二倍进行一次形态学滤波，第i次滤波结果为PN_i(f)；B2: The scale B_bi is twice the value of the previous filtering scale B_bi-1 to perform a morphological filtering, and the result of the i-th filtering is PN_i (f);

B3:将第i次和第i-1次相邻两次滤波结果差分，若ΔPN_i(f)＞ΔPN_i-1(f)，则将频点f处尺度更新为B_bi，否则尺度保持不变；B3: Difference between the i-th and i-1th adjacent filtering results, if ΔPN_i (f) > ΔPN_i-1 (f), then update the scale at the frequency point f to B_bi , otherwise the scale remains constant;

B4:若B_bi＝B_max，则停止迭代；B4: if B_bi =B_max , then stop iteration;

B5:各个频点f处以迭代确定的尺度B_bi(f)进行一次形态学滤波估计出有色噪声基底；B5: Carry out a morphological filter with the iteratively determined scale B_bi (f) at each frequency point f to estimate the colored noise floor;

B6:以原功率谱减B5估计出噪声基底得到噪底平坦的功率谱；B6: Estimate the noise floor with the original power spectrum minus B5 to obtain a flat power spectrum with a noise floor;

B7:将功率谱转换为对数功率谱，功率带个数越多，白噪声门限估计越准确，相邻功率带频点数分别差分，以差分结果最大的功率带的上限作为白噪声功率门限。B7: Convert the power spectrum to a logarithmic power spectrum. The more the number of power bands, the more accurate the white noise threshold estimation is. The frequency points of adjacent power bands are respectively differentiated, and the upper limit of the power band with the largest difference result is used as the white noise power threshold.

本实施例中，所述步骤D中具体包括如下步骤：In this embodiment, the step D specifically includes the following steps:

C1:在功率谱中寻找最大值位置f_max作为中心频率f_Ctr，在中心频率幅度P(f_Ctr)-3dB处计算出3dB带宽B_3dB，带宽左边界为f_3dBL，右边界为f_3dBR；C1: Find the maximum position f_max in the power spectrum as the center frequency f_Ctr , calculate the 3dB bandwidth B_3dB at the center frequency amplitude P(f_Ctr )-3dB, the left boundary of the bandwidth is f_3dBL , and the right boundary is f_3dBR ;

C2:将功率谱在[f_3dBL,f_3dBR]以尺度λ_sm＝0.1*B_3dB进行平滑滤波处理后返回C1，循环五次后进入C3；C2: return the power spectrum to C1 after smoothing and filtering the power spectrum at [f_3dBL , f_3dBR ] with the scale λ_sm =0.1*B_3dB , and enter C3 after five cycles;

C3:经过五次循环平滑后，所得带宽即为3dB带宽，在最终的[f_3dBL,f_3dBR]范围内根据下式计算中心频率；C3: After five cycles of smoothing, the resulting bandwidth is the 3dB bandwidth, and the center frequency is calculated according to the following formula within the final [f_3dBL , f_3dBR ] range;

C4:在f_3dBL以左，f_3dBR以右寻找功率谱首个极小值点或零点作为信号主瓣边界，左边界为f_LS，右边界为f_RS，在[f_LS,f_RS]范围内将信号归零剔除，根据下式计算带外斜率k_LS、k_RS：C4: To the left of f_3dBL and to the right of f_3dBR , look for the first minimum point or zero point of the power spectrum as the signal main lobe boundary, the left boundary is f_LS , the right boundary is f_RS , in the range of [f_LS , f_RS ] The signal is returned to zero and eliminated, and the out-of-band slopes k_LS and k_RS are calculated according to the following formula:

C5:在f_LS，f_RS处分别以斜率k_LS、k_RS向已估计信号带内[f_LS,f_RS]延伸，至两线相交为止，以两条延长线代替已估计信号的带内谱线，减小对为未估计信号的损伤；C5: At f_LS and f_RS , extend to the estimated signal band [f_LS , f_RS ] with slopes k_LS and k_RS respectively until the two lines intersect, and replace the estimated signal band with two extension lines Spectral lines, reducing damage to unestimated signals;

C6:若C4已估计能量低于最小判定能量或带宽小于最小检测带宽，则将其舍弃，若残余功率谱最大幅度不高于白噪声门限，则信号检测完成。C6: If the energy estimated by C4 is lower than the minimum decision energy or the bandwidth is smaller than the minimum detection bandwidth, it is discarded. If the maximum amplitude of the residual power spectrum is not higher than the white noise threshold, the signal detection is completed.

实施例五Embodiment five

如图1-7所示，一种基于尺度迭代和频谱补偿的宽带信号检测方法，包括如下步骤：As shown in Figure 1-7, a wideband signal detection method based on scale iteration and spectrum compensation includes the following steps:

步骤A:接收信号经过AD采样得到宽带采样信号，经过数字信道化将宽带信号划分为合适带宽的多段信号；Step A: the received signal obtains a broadband sampling signal through AD sampling, and divides the broadband signal into multi-section signals of suitable bandwidth through digital channelization;

步骤B:取数段宽带信号，据各段信号功率谱密度计算最大功率谱；Step B: take several sections of broadband signals, and calculate the maximum power spectrum according to the power spectral density of each section of signals;

步骤C:利用尺度迭代形态学滤波算法估计并去除有色噪声基底，并计算出白噪声功率门限；Step C: Utilize the scale iterative morphological filtering algorithm to estimate and remove the colored noise base, and calculate the white noise power threshold;

步骤D:检测功率谱中幅度最高的信号，估计出其中心频率、带宽以及信号频带起止位置，根据两次信号检测前后残余信号能量差值高于预设能量门限，则以补偿后的功率谱再次进行步骤D，否则信号检测结束。Step D: Detect the signal with the highest amplitude in the power spectrum, estimate its center frequency, bandwidth, and the start and end positions of the signal frequency band, and use the compensated power spectrum Perform step D again, otherwise the signal detection ends.

本实施例中，所述步骤A具体包括如下步骤：In this embodiment, the step A specifically includes the following steps:

A1：计算实信号信道化子信道的中心频率

整个频带被分为D个实信号的对称子带；A1: Calculate the center frequency of the channelized sub-channel of the real signal

The entire frequency band is divided into D symmetric subbands of real signals;

A2：抗混叠滤波器h_LP(n)为低通FIR滤波器,N与D具有整数倍关系,即K＝FD,则传统信道化结构中的第k个子信道的输出为：A2: The anti-aliasing filter h_LP (n) is a low-pass FIR filter, N and D have an integer multiple relationship, that is, K=FD, then the output of the kth sub-channel in the traditional channelization structure is:

A3：将上式改写为多相滤波结构，得到第k个子信道的输出表达式为：A3: Rewrite the above formula into a polyphase filter structure, and the output expression of the kth sub-channel is obtained as:

A4：将中心频率ω_k代入式中,此时第k个子信道的输出表达式为：A4: Substitute the center frequency ω_k into the formula, and the output expression of the kth sub-channel at this time is:

本实施例中，所述步骤B中，将接收数据进行m次分段，每段长度为L，段间间隔为K，提取数据可以矩阵形式表示：In this embodiment, in the step B, the received data is segmented m times, each segment length is L, and the interval between segments is K, and the extracted data can be expressed in matrix form:

对每段数据以周期图法估计出各段数据的功率谱P_i(f)，即For each piece of data, the power spectrum P_i (f) of each piece of data is estimated by the periodogram method, namely

其中，w(n)是所选窗函数，求出各分段信号功率谱后在每个频率点处求出功率谱的最大值，即Among them, w(n) is the selected window function, after obtaining the power spectrum of each segmented signal, the maximum value of the power spectrum is obtained at each frequency point, that is

本实施例中，所述步骤C具体包括如下步骤：In this embodiment, the step C specifically includes the following steps:

B1:尺度B_bi取值最小检测带宽B_min为进行一次形态学滤波PN₁(f)；B1: The minimum detection bandwidth B_min of the scale B_bi is to perform a morphological filter PN₁ (f);

B2:尺度B_bi取值上一次滤波尺度B_bi-1的二倍进行一次形态学滤波，以B_max为尺度进行一次形态学滤波，第i次滤波结果为PN_i(f)；B2: The value of the scale B_bi is twice the value of the previous filtering scale B_bi-1 to perform a morphological filter, and to perform a morphological filter on the scale of B_max , and the i-th filtering result is PN_i (f);

B3:将第i次和第i-1次相邻两次滤波结果差分，若ΔPN_i(f)＞ΔPN_i-1(f)，则将频点f处尺度更新为B_bi，否则尺度保持不变；B3: Difference between the i-th and i-1th adjacent filtering results, if ΔPN_i (f) > ΔPN_i-1 (f), then update the scale at the frequency point f to B_bi , otherwise the scale remains constant;

B4:若B_bi＝B_max，则停止迭代；B4: if B_bi =B_max , then stop iteration;

B5:各个频点f处以迭代确定的尺度B_bi(f)进行一次形态学滤波估计出有色噪声基底；B5: Carry out a morphological filter with the iteratively determined scale B_bi (f) at each frequency point f to estimate the colored noise floor;

B6:以原功率谱减B5估计出噪声基底得到噪底平坦的功率谱；B6: Estimate the noise floor with the original power spectrum minus B5 to obtain a flat power spectrum with a noise floor;

B7:在功率谱上由0至最大值划定适当个数的功率带并统计落在各带内的频点数，白噪声门限估计越准确，以差分结果最大的功率带的上限作为白噪声功率门限。B7: Delineate an appropriate number of power bands from 0 to the maximum value on the power spectrum and count the number of frequency points falling in each band. The white noise threshold estimation is more accurate, and the upper limit of the power band with the largest difference result is used as the white noise power. threshold.

本实施例中，所述步骤D中具体包括如下步骤：In this embodiment, the step D specifically includes the following steps:

C1:在功率谱中寻找最大值位置f_max作为中心频率f_Ctr，在中心频率幅度P(f_Ctr)-3dB处计算出3dB带宽B_3dB，带宽左边界为f_3dBL，右边界为f_3dBR；C1: Find the maximum position f_max in the power spectrum as the center frequency f_Ctr , calculate the 3dB bandwidth B_3dB at the center frequency amplitude P(f_Ctr )-3dB, the left boundary of the bandwidth is f_3dBL , and the right boundary is f_3dBR ;

C2:将功率谱在[f_3dBL,f_3dBR]以尺度λ_sm＝0.1*B_3dB进行平滑滤波处理后返回C1，循环五次后进入C3；C2: return the power spectrum to C1 after smoothing and filtering the power spectrum at [f_3dBL , f_3dBR ] with the scale λ_sm =0.1*B_3dB , and enter C3 after five cycles;

C3:经过五次循环平滑后，所得带宽即为3dB带宽，在最终的[f_3dBL,f_3dBR]范围内根据下式计算中心频率；C3: After five cycles of smoothing, the resulting bandwidth is the 3dB bandwidth, and the center frequency is calculated according to the following formula within the final [f_3dBL , f_3dBR ] range;

C4:在f_3dBL以左，f_3dBR以右寻找功率谱首个极小值点或零点作为信号主瓣边界，左边界为f_LS，右边界为f_RS，在[f_LS,f_RS]范围内将信号归零剔除，根据下式计算带外斜率k_LS、k_RS：C4: To the left of f_3dBL and to the right of f_3dBR , look for the first minimum point or zero point of the power spectrum as the signal main lobe boundary, the left boundary is f_LS , the right boundary is f_RS , in the range of [f_LS , f_RS ] The signal is returned to zero and eliminated, and the out-of-band slopes k_LS and k_RS are calculated according to the following formula:

C5:在f_LS，f_RS处分别以斜率k_LS、k_RS向已估计信号带内[f_LS,f_RS]延伸，至两线相交为止，减小对为未估计信号的损伤；C5: At f_LS and f_RS , extend to the estimated signal band [f_LS , f_RS ] with slopes k_LS and k_RS respectively until the two lines intersect, reducing the damage to the unestimated signal;

C6:若C4已估计能量低于最小判定能量或带宽小于最小检测带宽，则将其舍弃，若残余功率谱最大幅度不高于白噪声门限，则信号检测完成。C6: If the energy estimated by C4 is lower than the minimum decision energy or the bandwidth is smaller than the minimum detection bandwidth, it is discarded. If the maximum amplitude of the residual power spectrum is not higher than the white noise threshold, the signal detection is completed.

由此，信号检测环节流程如图4所示。Therefore, the process flow of the signal detection link is shown in Figure 4 .

算法性能仿真结果如图5-图7所示，仿真参数设定：信号采样率为8MHz，符号速率为2MBaud，调制方式为QPSK。采样时间设为25ms，共有3个TDMA信号，每个信号突发时长为0.36ms。信道环境为加性高斯白噪声。The algorithm performance simulation results are shown in Figures 5-7. The simulation parameters are set: the signal sampling rate is 8MHz, the symbol rate is 2MBaud, and the modulation method is QPSK. The sampling time is set to 25ms, there are 3 TDMA signals, and the burst duration of each signal is 0.36ms. The channel environment is additive white Gaussian noise.

为验证本专利的有色噪声基底估计准确度，图5给出了某宽带功率谱有色噪声基底估计结果及剔除噪底后效果。信号带宽37.5MHz，中心频率为963MHz，采样率为75MHz。可以看出，噪声基底同信号起伏贴合紧密，去除有色噪声后功率谱噪底平坦，仅残余白噪声，且信号个数无变化，说明该算法具有良好的噪底估计性能，适合在起伏噪底条件下使用。In order to verify the accuracy of the colored noise floor estimation of this patent, Fig. 5 shows the estimation results of the colored noise floor of a broadband power spectrum and the effect after removing the noise floor. The signal bandwidth is 37.5MHz, the center frequency is 963MHz, and the sampling rate is 75MHz. It can be seen that the noise floor fits closely with the signal fluctuations, and the noise floor of the power spectrum is flat after removing the colored noise, only residual white noise, and the number of signals does not change, which shows that the algorithm has good noise floor estimation performance and is suitable for fluctuation noise Use under low conditions.

为验证本专利的有多信号检测性能，图6给出了某宽带功率谱信号检测结果。可以看出，强弱信号均可检出，且当信号重叠时，强信号并未对弱信号的检出造成影响。In order to verify the multi-signal detection performance of this patent, Figure 6 shows the detection results of a certain broadband power spectrum signal. It can be seen that both strong and weak signals can be detected, and when the signals overlap, the strong signal does not affect the detection of weak signals.

为进一步验证本文算法性能的稳定性，设置实验条件为：宽带信号内有25个子信号，调制方式包括BPSK、QPSK、8PSK、8QAM、16QAM，信号带宽范围为2k～5MHz，高斯白噪声信道，信噪比为-14～6dB。由于能量检测算法、梯度双阈值算法以及功率谱对消法均受不平坦噪底影响较大，功率谱噪底均为平坦噪底。如图7可以看出，本文算法正确检出概率始终远高于经典的能量检测算法，在信噪比极低时同梯度双阈值算法以及功率谱对消法相差不大，在-10dB开始高于三种算法，在SNR≥0dB时检测概率高于90％。In order to further verify the stability of the algorithm performance in this paper, the experimental conditions are set as follows: there are 25 sub-signals in the broadband signal, the modulation methods include BPSK, QPSK, 8PSK, 8QAM, 16QAM, the signal bandwidth ranges from 2k to 5MHz, Gaussian white noise channel, signal The noise ratio is -14~6dB. Since the energy detection algorithm, the gradient double threshold algorithm and the power spectrum cancellation method are greatly affected by the uneven noise floor, the power spectrum noise floor is flat. As shown in Figure 7, the correct detection probability of the algorithm in this paper is always much higher than that of the classical energy detection algorithm. When the signal-to-noise ratio is extremely low, it is not much different from the gradient double-threshold algorithm and the power spectrum cancellation method. It starts to be higher at -10dB For the three algorithms, the detection probability is higher than 90% when SNR≥0dB.

具体的基于尺度迭代和频谱补偿的宽带信号检测与测量方法系统计算过程可参见上述实施例，本发明实施例在此不再赘述。For the specific calculation process of the broadband signal detection and measurement method system based on scale iteration and spectrum compensation, refer to the above-mentioned embodiments, and the embodiments of the present invention will not be repeated here.

最后应说明的是：以上实施例仅用以说明本发明的技术方案，而非对其限制；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换，但这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的精神和范围。Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features, but these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (6)

1. A broadband signal detection and measurement method based on scale iteration and spectrum compensation is characterized by comprising the following steps:

a, receiving signals are subjected to AD sampling to obtain broadband sampling signals, and the broadband signals are divided into multiple sections of signals with proper bandwidths through digital channelization;

b, taking several sections of broadband signals, and calculating a maximum power spectrum according to the power spectral density of each section of signal;

estimating and removing a colored noise substrate by using a scale iteration morphological filtering algorithm, and calculating a white noise power threshold;

and D, detecting the signal with the highest amplitude in the power spectrum, estimating the center frequency, the bandwidth and the starting and stopping positions of the signal frequency band, eliminating the signal according to a compensation algorithm, finishing compensation, detecting one signal each time, performing the step D again according to the compensated power spectrum if the energy difference value of the residual signal before and after the signal detection for two times is higher than a preset energy threshold, and otherwise, finishing the signal detection.

2. The wideband signal detection method based on scale iteration and spectrum compensation as claimed in claim 1, wherein the characteristic of morphological filtering bandwidth screening is utilized to perform morphological filtering iteration on the power spectrum, so that the noise floor estimation is more accurate, the error caused by strong pulse and white noise in the signal to the signal parameter estimation is overcome by utilizing smooth iteration, and the detection performance of the band-aliased signal is improved by spectrum compensation.

3. The wideband signal detection method based on scale iteration and spectral compensation according to claim 2, wherein the step a specifically includes the steps of:

a1: calculating center frequencies of real signal channelized sub-channels

D is a data extraction multiple, and the whole frequency band is divided into D symmetrical sub-bands of real signals;

a2: anti-aliasing filter h_LP (N) is the low-pass FIR filter, K is the number of channels, D is the decimation factor, and N and D have an integer multiple relationship, i.e., K = FD, then the output of the kth sub-channel in the conventional channelization structure is:

a3: the above formula is rewritten as a polyphase filter structure, and the output expression of the k-th sub-channel is obtained as follows:

a4: centering frequency omega_k In the formula, the output expression of the kth sub-channel at this time is:

4. the wideband signal detection method based on scale iteration and spectrum compensation according to claim 3, wherein in step B, the received data is segmented m times, each segment has a length of L and an interval of K, and the extracted data can be represented in a matrix form:

estimating the power spectrum P of each segment of data by a periodogram method_i (f) I.e. by

Where w (n) is a selected window function, the power spectrum of each segmented signal is found, and the maximum value of the power spectrum is found at each frequency point, i.e.

5. The wideband signal detection method based on scale iteration and spectral compensation according to claim 4, wherein the step C specifically comprises the steps of:

b1: dimension B_bi Dereferencing minimum detection bandwidth B_min For performing morphological filtering PN once₁ (f)；

B2 is the scale B_bi Last filtering scale B_bi-1 Twice as many as the maximum detection bandwidth B, if the size is larger than the maximum detection bandwidth_max Then with B_max Performing morphological filtering for scale, wherein the ith filtering result is PN_i (f)；

B3, differentiating the filtering results of the ith and the (i-1) th adjacent two times to obtain a differential result delta PN_i Comparing the difference results twice frequency point by frequency point if delta PN_i (f)＞ΔPN_i-1 (f) Then the scale at the frequency point f is updated to B_bi Otherwise, the dimension is kept unchanged;

b4 if B_bi ＝B_max Then stopping iteration;

b5, determining the dimension B at each frequency point f by iteration_bi (f) Performing primary morphological filtering to estimate a colored noise substrate;

b6, subtracting B5 from the original power spectrum to estimate a noise base to obtain a power spectrum with flat noise base;

and B7, converting the power spectrum into a logarithmic power spectrum, defining a proper number of power bands from 0 to the maximum value on the power spectrum, counting the number of frequency points falling in each band, wherein the more the number of the power bands is, the more accurate the white noise threshold estimation is, the frequency point numbers of adjacent power bands are respectively differentiated, and the upper limit of the power band with the maximum differential result is used as the white noise power threshold.

6. The wideband signal detection and measurement method based on scale iteration and spectral compensation according to claim 5, wherein the step D specifically includes the following steps:

c1 finding the position f of the maximum in the power spectrum_max As a center frequency f_Ctr Amplitude P (f) at the center frequency_Ctr ) -3dB calculation of the 3dB bandwidth B_3dB Left boundary of bandwidth is f_3dBL The right boundary is f_3dBR ；

C2, the power spectrum is in f_3dBL ,f_3dBR ]In the dimension λ_sm ＝0.1*B_3dB Performing smoothing filteringReturning to C1 after treatment, and entering C3 after five times of circulation;

c3, after five times of circular smoothing, the obtained bandwidth is 3dB bandwidth, and the final bandwidth is f_3dBL ,f_3dBR ]Calculating the center frequency within the range according to the following formula;

c4 at f_3dBL To the left, f_3dBR The first minimum value point or zero point of the right-sought power spectrum is used as the boundary of the main lobe of the signal, and the left boundary is f_LS The right boundary is f_RS In [ f_LS ,f_RS ]Within the range, the signal is zeroed and eliminated, and the out-of-band slope k is calculated according to the following formula_LS 、k_RS ：

C5 at f_LS ，f_RS At a slope k_LS 、k_RS Into the estimated signal band f_LS ,f_RS ]Extending until the two lines intersect, and replacing the in-band spectral line of the estimated signal by the two extension lines to reduce the damage to the estimated signal;

and C6, if the estimated energy of the C4 is lower than the minimum judgment energy or the bandwidth is smaller than the minimum detection bandwidth, discarding the estimated energy or the bandwidth, and if the maximum amplitude of the residual power spectrum is not higher than a white noise threshold, finishing the signal detection.

Images