CN103634026B - Digital zero intermediate frequency self-adaptation wave trapping method based on FPGA (filed programmable gate array) - Google Patents

Abstract

The invention discloses a novel digital zero-IF self-adaptive trapping method based on FPGA, which comprises the following steps: Step 1: Transform the sampled IF signal into the frequency domain by using FFT transformation; Step 2: Perform spectrum statistics on the IF signal in the frequency domain Analysis and threshold judgment, judge the number of NBI and its center frequency; step 3: if there is no interference, go to step 7; if there is interference, carry out quadrature demodulation to the sampled intermediate frequency signal through the quadrature digital down converter, and Use the low-pass filter to retain only the positive frequency component of the sampled intermediate frequency band-pass signal, and transform to obtain the baseband signal; Step 4: Perform spectrum shift on the baseband signal, move the first interference point to zero frequency, and filter out the interference through the high-pass filter Spectrum components, thereby suppressing the first interference point; Step 5: Repeat the operation of Step 4 until all other interference points are suppressed; Step 6: Transform the signal after interference suppression processing into a real signal; Step 7: The baseband processing module to demodulate the output.

Description

A FPGA-Based Digital Zero-IF Adaptive Notch Method

technical field

The invention belongs to the technical field of communication, and relates to strong narrow-band interference suppression in a spread spectrum communication system, in particular to an FPGA-based digital zero-IF self-adaptive trap.

Background technique

In the direct spread communication system (DDDS), a typical low probability of intercept signal is transmitted, the signal occupies a wide spectrum, and the power density is low. It has good suppression performance for various types of interference, but in practice it is often limited by the system bandwidth. The inherent spread spectrum gain of the system itself is not enough to suppress the high-power strong narrow-band interference NBI (Narrow-Band Interference). Therefore, it is necessary to use signal processing technology to suppress strong narrowband interference without increasing the signal bandwidth before despreading, so as to improve the anti-interference ability of the system.

At present, the NBI suppression technology of the spread spectrum system is mainly divided into three types: time domain interference suppression technology, transform domain interference suppression technology, and code-assisted interference suppression technology. Regarding these three technologies, Guo Lili, Yin Fulian, Lu Manhong and others published their advantages and disadvantages in the article "Overview of DSSS/CDMA System Narrowband Interference Suppression Technology [J]" (Acta Electronics. 2009,10:2248-2257). Introduced in detail, the common feature of these three technologies is to design a variety of notch filters with different algorithms, first transform the intermediate frequency signal in various ways, then detect the NBI in the signal, and use the notch filter The filter eliminates the specific frequency NBI in the signal, and finally restores and compensates the intermediate frequency signal to facilitate subsequent signal baseband processing and demodulation and despreading. Figure 1 shows the traditional NBI suppression model of the spread spectrum system based on these three technologies. It can be seen from the figure that this method closely combines interference detection and interference suppression. Before NBI detection, the original signal of the intermediate frequency x (t) do preprocessing, no matter whether the system has NBI or not, it will damage x(t) to a certain extent. Although various methods are used to compensate and restore the intermediate frequency signal in the later stage, it increases the complexity of system design and hardware overhead. Difficult to implement in engineering.

Contents of the invention

Therefore, for the above-mentioned problems, the present invention proposes a novel digital zero-IF adaptive notch filter based on FPGA, which simplifies the design algorithm, reduces the complexity and hardware overhead of system design, and effectively suppresses the NBI in the spread spectrum system, thereby solving the problem of Insufficiency of existing technology.

In order to solve the above-mentioned technical problems, the design concept adopted in the present invention is to completely separate the interference detection and interference suppression modules, and the task of the detection module is only to use traditional NBI detection technology to judge whether there is NBI in the spread spectrum system. If there is no NBI, directly Output the original intermediate frequency signal x(t); if there is NBI, copy the x(t) signal, and then send it to the interference suppression module to suppress the NBI notch and output x'(t). In this way, since there is no need to perform various recovery compensations on the signal processed by the interference detection, the complexity of system implementation is greatly reduced and the reliability of the system is improved. The present invention adopts above-mentioned new spread spectrum system NBI suppressing model, combines the digital zero-IF processing technology of programmable logic gate array (FPGA), designs a kind of novel FPGA digital zero-IF self-adaptive notch method and notch filter. The notch method first performs frequency-domain NBI detection based on threshold estimation on the sampled digital IF signal, and analyzes NBI statistics to find the normalized center frequency of NBI; then uses digital quadrature demodulation to move the IF signal to zero-IF (baseband), and move the NBI through the spectrum to zero-frequency notch filtering in the baseband, and then reconstruct the signal to restore the IF bandpass signal (real signal), and finally send it to the baseband processing module for demodulation output. At the same time, the two main modules of the notch filter are given: the frequency domain NBI detection module based on threshold estimation and the specific implementation method of the digital zero-IF NBI suppression module, to realize the adaptive trapping of NBI in the direct sequence spread spectrum system (DSSS) wave suppression.

Specifically, a novel digital zero-IF self-adaptive notch method based on FPGA of the present invention comprises the following steps:

Step 1: Transform the sampled intermediate frequency signal into the frequency domain by FFT transform;

Step 2: Perform spectrum statistics analysis and threshold judgment on the intermediate frequency signal in the frequency domain, and judge the number of NBIs and their center frequency;

Step 3: If there is no interference, go to step 7 to directly output the sampled original IF signal; if there is interference, perform interference suppression, and perform quadrature demodulation on the sampled IF signal through a quadrature digital down-converter (DDC) , and use the low-pass filter to retain only the positive frequency component of the sampled intermediate frequency band-pass signal, and transform to obtain the baseband signal;

Step 4: Spectrum shifting is performed on the baseband signal, the first interference point is moved to zero frequency, and the interference spectrum component is filtered out through a high-pass filter, thereby suppressing the first interference point;

Step 5: If there is more than one interference point, repeat the operation of step 4, that is, the signal after the first or second interference suppression processing is then shifted to the spectrum, the interference point is moved to zero frequency, and the interference is filtered out through a high-pass filter Spectral components, thereby suppressing other interference points (second or third interference points, etc.);

Step 6: Transform the signal after the interference suppression processing into a real signal (intermediate frequency bandpass signal);

Step 7: Demodulate and output by the baseband processing module.

Wherein, in the step 2, the threshold judgment adopts the NBI threshold detection and estimation method based on the transform frequency domain. In the prior art, commonly used methods for designing the interference detection threshold include the N-sigma algorithm, the σ ² maximum likelihood estimation method, and the median filter method. Since the N-sigma algorithm is relatively easy to implement in engineering, the present invention adopts this algorithm and performs certain engineering approximate calculations on the relevant parameters of the algorithm. Specifically, the threshold judgment includes the following steps:

Step 21: Establish a signal model. For a direct spread communication system, the signal is spread on a wide frequency spectrum by multiplying the signal with a pseudo-random (Pseudo-Noise, PN) sequence at the transmitting end; at the receiving end, by multiplying the same The PN sequence is multiplied with the received signal to recover the signal. The received intermediate frequency signal x(t) generally consists of 3 parts:

x(t)=s(t)+i(t)+n(t) (1)

Where s(t) is the spread spectrum signal, i(t) is the narrowband interference, and n(t) is the additive Gaussian white noise.

After x(t) is sampled at the PN symbol rate, it can be expressed as:

x(k)=s(k)+i(k)+n(k) (2)

Among them, s(k) and n(k) are not related in time, but have correlation with i(k);

Step 22: NBI suppression first needs to sample the intermediate frequency signal, and perform spectrum analysis on the sampled digital signal through a 1024-point FFT to the received signal. Through statistical analysis of the spectrum information, find out the interference point, and then carry out NBI suppression. Its goal is to remove narrow-band interference from signals over a wide frequency range. These disturbances occupy fewer frequency bins and are higher in magnitude than the noise floor of the FFT. In the absence of narrowband interference, the spread spectrum signal plus noise is an approximate Gaussian distribution. As interference is superimposed on the signal, the spectrum distribution is distorted. In the part that is not affected by interference, the spectrum distribution of the signal maintains the shape of Gaussian distribution; in the frequency part that is affected, the spectrum distribution is above the desired signal spectrum. In the absence of disturbances, most signals are smaller than the mean (μ) plus twice the standard deviation (σ); when there is a series of disturbances or the total energy of the disturbances increases, the mean rises relative to the mean of the part of the Gaussian distribution, and The standard deviation also increases due to the presence of noise. This distortion requires selecting a threshold closer to the mean to filter out the frequency components associated with the interfering signal. Therefore, the determination method of the filtering threshold of the present invention is as follows: first, the FFT data is divided into different data blocks according to every 1024 points, and for each 1024-point FFT data block, the decibel value of each frequency point amplitude in the data block is calculated , and then calculate the standard deviation σ and mean value μ of the data block; select an appropriate weighting factor N from the preset weighted set according to the standard deviation σ, and set the interference detection gate as:

Th＝μ+Nσ (3)

In the formula, N is the weighting factor;

σ is divided into 5 levels σ ₀ , σ ₁ , σ ₂ , σ ₃ , σ ₄ , N is also divided into N ₀ , N ₁ , N ₂ , N ₃ , N ₄ , and the choice of N is related to σ.

$\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\\ {\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\\ {\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

The preset weighting factor set can be adjusted according to different channel environments, so that after the signal undergoes FFT, N is selected according to the mean value and standard deviation to obtain the interference detection threshold. In the N-sigma algorithm, the determination and selection criteria of N are the key to the algorithm design. If the value of N is too large, the estimation threshold will be too high, causing serious interference leakage; otherwise, if the value of N is too small, the estimation threshold will be too low, resulting in misjudgment.

Larger σ indicates stronger interference, and a small scaling factor is needed to keep the gate on top of and close to the noise. Each frequency point is compared with the threshold one by one, the area greater than the threshold is recorded and its center position is found, and the center position is the center frequency of the normalized interference. Since the values of the five candidate proportional coefficients (N ₀ -N ₄ ) are determined by FPGA programming, the algorithm can be easily applied to other system applications.

Step 23: Engineering Approximate Calculation of Statistics of Variance and Mean: The interference detection analysis calculates σ and μ for each FFT data block, and these statistics are used to calculate the threshold. In order to calculate these statistics, the output decibel value of each FFT data block is 10log ₁₀ (|X(k)|), where |X(k)| is the amplitude of the sampled intermediate frequency signal X(k), and the logarithm is used here The coordinates are used to reduce the spectral distortion of the signal mixed by spread spectrum signal, thermal noise and interference, and at the same time reduce the numerical accuracy requirements of |X(k)| required for the subsequent calculation of statistics. In order to reduce the hardware implementation Complexity, with some approximations to the computation of 10log ₁₀ (|X(k)|) and statistics. Through simulation and actual verification, the accuracy drop caused by approximate calculation does not affect the identification of interference signals from spread spectrum signals and thermal noise mixed signals. The amplitude |X(k)| of the sampled intermediate frequency signal X(k) is approximately calculated as follows:

|X(k)|≈[max(ReX(k), ImX(k))+min(ReX(k), ImX(k))/4]    (5)

Wherein Re is the real part of the signal, Im is the imaginary part of the signal, the overall error of the above approximate operation is 0.6%, and the maximum error is 11.6% (at π/4, 3π/4, 5π/4, 7π/4 respectively).

The approximate calculation of the logarithmic quantization is as follows:

10log ₁₀ (|X(k)|)＝10log ₁₀ (2)·log ₂ (|X(k)|)≈3log ₂ (|X(k)|) (6)

It can be seen that the magnitude calculation can be completed only by shifting and adding operations, while avoiding multiplication and square root operations that are sensitive to hardware resources. The overall error of the above logarithmic operations is 0.5%, and the maximum error is 2.9%.

The mean and standard deviation of each FFT data block are calculated as follows:

$\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}{\mathrm{<span\; class="notranslate">\; log</span>}}_{\mathrm{<span\; class="notranslate">\; 10</span>}}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

${\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}}<span\; class="notranslate">\; [</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\mathrm{<span\; class="notranslate">\; 10</span>}}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}}<span\; class="notranslate">\; (</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\mathrm{<span\; class="notranslate">\; 10</span>}}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

Wherein, _NP is the number of frequency points of each FFT data block. According to the above equation, the accumulator accumulates 10log ₁₀ (|X(k)|) and (10log ₁₀ (|X(k)|)) ² of N _P frequency points. After each block is processed, the accumulated value is used for Calculate the mean value μ and variance σ ² , σ can be derived from the square root of σ ² .

Further, the signal suppression in step 3 includes the following steps:

Step 31, construct the baseband signal: make the mathematical expression of the intermediate frequency bandpass signal be:

$\begin{array}{c}\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ <span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>}<span\; class="notranslate">\; +</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>}\end{array}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 9</span>}<span\; class="notranslate">\; )</span>$

where a(t) is the information sequence, f ₀ is the center frequency of the carrier, is the carrier phase, then the mathematical expression of the frequency shift filtered baseband signal S _B (t) is:

In the formula S _BI (t) and S _BQ (t) are respectively the mid

The same direction component and quadrature component (I component and Q component) of the frequency bandpass signal, is the frequency shift factor;

Step 32: Suppress interference points:

The interference component contained in the baseband signal S _B (t) is recorded as f _I , then the baseband signal is multiplied by the frequency shift factor e ^-j2πfIt to obtain the baseband signal S _B ^* (t) where f _I is moved to zero frequency, S _B ^* (t) After high-pass filtering, the baseband signal S _B '(t) that suppresses the interference f _I is obtained; the calculation formula of the baseband signal S _B '(t) that suppresses the interference f _I is as follows:

(S _I +jS _Q )×[cos(ω ₀ t)+jsin(ω ₀ t)]

＝[S _I ×cos(ω ₀ t)-S _Q ×sin(ω ₀ t)]+j[S _I ×sin(ω ₀ t)+S _Q ×cos(ω ₀ t)] (11)

Where ω ₀ =-2πf _I ;

Step 33: IF bandpass signal reconstruction:

The signal after interference suppression is a baseband signal (complex signal), which needs to be converted into a real signal for subsequent processing. By analyzing the intermediate frequency bandpass signal S(t), the analytical signal S _A (t), and the baseband signal S _B (t) The relationship between them can get the mathematical expression of reconstructing the real signal from the baseband signal. The analytical signal S _A (t) can be expressed as:

$\begin{array}{c}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; A</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\\ <span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; B</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{\mathrm{<span\; class="notranslate">\; j</span>}\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}}\\ <span\; class="notranslate">\; =</span><span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BI</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BQ</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ <span\; class="notranslate">\; =</span><span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BI</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BQ</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\\ <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BI</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BQ</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\end{array}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 12</span>}<span\; class="notranslate">\; )</span>$

in is the Hilbert transform of the intermediate frequency bandpass signal S(t), f _c is the center frequency of the intermediate frequency bandpass signal, and the bandwidth of the intermediate frequency bandpass signal is 2B. Comparing the two sides of equation (12), the required real signal is:

S(t) = S _BI (t) cos(2πf _c t) - S _BQ (t) sin(2πf _c t).

A novel digital zero-IF self-adaptive trap based on FPGA of the present invention comprises an analog-to-digital conversion unit (ADC), an FPGA unit, a configuration logic circuit, a memory and a user interface, and the intermediate frequency input signal is sent through the analog-to-digital conversion unit (ADC) To the FPGA unit, the memory and the user interface are connected to the configuration logic circuit in two-way communication, and the output terminal of the configuration logic circuit is connected to the input terminal of the FPGA unit; the FPGA unit includes an FFT (Fast Fourier Transform) module and a spectrum-based interference detection and analysis module , an interference adaptive notch suppression module and a baseband signal processing module, the output of the FFT module is connected to the input of the interference detection and analysis module, and the output of the interference detection and analysis module is connected to the input of the interference adaptive notch suppression module and the baseband The input end of the signal processing module and the output end of the interference adaptive notch suppression module are also connected to the input end of the baseband signal processing module.

The present invention provides a novel FPGA-based digital zero-IF input signal self-adaptive notch method, which uses the transformation domain threshold interference detection technology and the idea of digital quadrature demodulation to perform real-time sampling and spectrum analysis on the IF input signal, And through the program preset parameters, select the appropriate suppression threshold, move the signal to zero frequency for filtering, after NBI filtering, then restore the baseband signal to an intermediate frequency bandpass signal for demodulation and despreading, and realize NBI adaptive suppression of DDDS . The invention greatly simplifies the design algorithm, reduces the complexity of system design and hardware overhead, and effectively suppresses NBI in the spread spectrum system.

Description of drawings

Fig. 1 provides the traditional NBI suppression model based on the prior art of the spread spectrum system;

Fig. 2 is the NBI suppression model of spread spectrum system of the present invention;

Fig. 3 is FPGA digital zero intermediate frequency self-adaptive notch filter structure of the present invention;

Fig. 4 is the narrowband interference suppression process of the present invention;

Fig. 5 is the flowchart of constructing baseband signal by intermediate frequency bandpass signal of the present invention;

Fig. 6 a is that the intermediate frequency bandpass signal of the present invention constructs the baseband signal spectrum shifting process (the frequency spectrum of the intermediate frequency bandpass signal);

Fig. 6 b is that the intermediate frequency bandpass signal of the present invention constructs the baseband signal spectrum shifting process (the frequency spectrum of the intermediate frequency bandpass signal after the frequency shift);

Fig. 6 c is that the intermediate frequency band-pass signal of the present invention constructs the baseband signal spectrum shifting process (low-pass filter amplitude-frequency characteristic);

Fig. 6 d is the process of shifting the spectrum of the intermediate frequency bandpass signal of the present invention to construct the baseband signal (the intermediate frequency bandpass signal after the frequency shift generates a baseband signal through low-pass filtering);

Fig. 7 is the interference point suppression process of the present invention;

Fig. 8 a is the frequency spectrum of the interference suppression process of the present invention (there is interference f _I in the signal);

Fig. 8 b is the frequency spectrum of the interference suppression process of the present invention (moving the interference f ₁ to zero frequency);

Fig. 8c is the frequency spectrum (high-pass filter amplitude-frequency characteristic) of the interference suppression process of the present invention;

Fig. 8d is the frequency spectrum of the interference suppression process of the present invention (interference is suppressed after high-pass filtering);

Fig. 9 is the process of reconstructing the real signal from the baseband signal of the present invention;

Fig. 10a is the frequency spectrum of baseband signal S _B (t) among the present invention;

Fig. 10b is the frequency spectrum of the intermediate frequency bandpass signal S (t) among the present invention;

Fig. 10c is the frequency spectrum of the analysis signal S _A (t) in the present invention;

Figure 11 is the background noise without intermediate frequency signal input;

Fig. 12 is the simulation curve of the signal spectrum before suppression;

FIG. 13 is a simulation curve of the signal spectrum after notch suppression.

Detailed ways

The present invention will be further described in conjunction with the accompanying drawings and specific embodiments.

The present invention aims at the deficiencies of the existing narrow-band interference (NBI, Narrow-Band Interference) suppression method in the spread spectrum system, proposes a new method model of narrow-band interference suppression, and provides the digital zero-IF self-adaptive under this model The specific implementation method of the notch filter. Its idea comes from the quadrature demodulation in the communication system, which is different from the general processing technology in structure and algorithm. The simulation and actual measurement verification show that the direct sequence spread spectrum system (DSSS) using this notch filter method can have a good suppression effect on strong narrowband interference.

Aiming at the deficiency of this method, the present invention proposes a method for suppressing NBI by a novel spread spectrum system. Its design idea is to completely separate the interference detection and interference suppression modules. The task of the detection module is only to use the traditional NBI detection technology to judge whether the spread spectrum system has NBI. If there is no NBI, it will directly output the IF input signal x(t); if there is NBI copies the x(t) signal, and then sends it to the interference suppression module to suppress the NBI notch and output x'(t). Figure 2 shows the implementation model of this new method. Since it does not need to perform various restoration and compensation on the signal processed by interference detection, it greatly reduces the complexity of system implementation and improves the reliability of the system.

The present invention adopts this new spread spectrum system NBI suppression model, combines the digital zero-IF processing technology of programmable logic gate array (FPGA), designs a kind of novel FPGA digital zero-IF self-adaptive notch, and provides The two main modules of this notch filter: the specific implementation method of the frequency domain NBI detection module based on threshold estimation and the digital zero-IF NBI suppression module, to realize the adaptive notch suppression of NBI in the direct sequence spread spectrum system (DSSS) .

1 system model

With the development of FPGA hardware and software radio technology, the intermediate frequency digital processing technology of spread spectrum communication system is becoming more and more mature. The digital zero intermediate frequency adaptive notch filter designed by the present invention is based on the threshold estimation of the digital intermediate frequency input signal after sampling. NBI detection in the frequency domain, and statistical analysis of NBI, to find the normalized center frequency of NBI. Use digital quadrature demodulation to move the IF input signal to zero-IF (baseband), and move NBI through the spectrum to zero-frequency notch filter in the baseband, and then reconstruct the IF bandpass signal (real signal) through signal reconstruction. Finally, it is sent to the baseband processing module for demodulation output.

1.1 FPGA zero-IF digital adaptive notch filter model

Figure 3 shows the structure of FPGA-based digital zero-IF adaptive notch filter.

The digital zero-IF notch filter is composed of analog-to-digital converter (ADC), FPGA, configuration logic circuit, memory and user interface and other circuits. The FPGA is composed of a Fast Fourier Transform (FFT) module, a spectrum-based interference detection and analysis module, an interference adaptive notch suppression module, and a baseband signal processing module. The programs that realize various functions are stored in the memory, and the FPGA is dynamically configured by looking up the table to realize the statistical analysis of the current NBI signal spectrum, digital quadrature demodulation, NBI zero-IF notch suppression, band-pass signal reconstruction and Baseband demodulation output and other functions.

1.2 Interference detection and suppression method

Firstly, FFT transforms the sampled IF input signal to the frequency domain, and performs spectrum statistics analysis and threshold judgment on the IF input signal in the frequency domain to determine the number of NBIs and their center frequency. If there is no interference, the sampled original IF input signal is directly demodulated and output by the baseband processing module. In the case of interference, the sampled IF input signal is demodulated by a quadrature digital down-converter (DDC), and only the positive frequency component of the sampled IF bandpass signal is retained through a low-pass filter, and transformed into a baseband signal (zero-IF input signal).

The baseband signal performs spectrum shifting, moves the first interference point to zero frequency, and filters out the interference spectrum components through a high-pass filter, thereby suppressing the first interference point. If there is more than one interference point, the signal after the first or second interference suppression processing is then spectrum shifted, the interference point is moved to zero frequency, and the interference spectrum component is filtered out through a high-pass filter, thereby suppressing the second or a third point of interference. Finally, the signal processed by the interference suppression is transformed into a real signal (intermediate frequency band-pass signal), which is demodulated and output by the baseband processing module. Its implementation process is shown in Figure 4.

2 Design of NBI detection module based on threshold judgment

To avoid NBI, accurate channel estimation is the basis for effective communication of the system. The detection of the electromagnetic environment can be accomplished by power spectrum estimation in the transform domain. After years of development of spectral estimation technology, there are many mature methods, such as periodogram method, auto regressive (AR) method and wavelet/wavelet packet spectral estimation technology. By comparing the power spectrum obtained by spectrum estimation with the interference detection threshold, the location of the interference can be determined. For the design of the interference detection threshold, commonly used methods include N-sigma algorithm, σ ² maximum likelihood estimation method, adaptive multi-threshold method, and median filter method. The design of the present invention adopts the N-sigma algorithm which is relatively easy to realize in engineering, and certain engineering application improvement is carried out on the algorithm.

2.1 Signal model

The present invention mainly studies the direct spread communication system (DDDS), and spreads the signal on a wide frequency spectrum by multiplying the signal with a pseudo-random (Pseudo-Noise, PN) sequence at the transmitting end. At the receiving end, the signal is recovered by multiplying the same PN sequence with the received signal. The intermediate frequency input signal x(t) received by DDDS generally consists of 3 parts:

x(t)=s(t)+i(t)+n(t) (1)

Where s(t) is a spread spectrum signal; i(t) is narrowband interference; n(t) is additive Gaussian white noise.

After x(t) is sampled at the PN chip rate, it can be expressed as:

x(k)=s(k)+i(k)+n(k) (2)

Among them, s(k) and n(k) are not related in time, but have correlation with i(k).

2.2 Detection method

NBI suppression first needs to sample the IF input signal, and perform spectrum analysis on the received signal through 1024-point FFT on the sampled digital signal. Through statistical analysis of the spectrum information, find out the interference point, and then carry out NBI suppression.

The goal of NBI suppression is to remove narrowband interference from signals over a wide frequency range. These disturbances occupy fewer frequency bins and are higher in magnitude than the noise floor of the FFT. In the absence of narrowband interference, the spread spectrum signal plus noise is an approximate Gaussian distribution. As interference is superimposed on the signal, the spectrum distribution is distorted. In the part that is not affected by interference, the spectrum distribution of the signal maintains the shape of Gaussian distribution; in the frequency part that is affected, the spectrum distribution is above the desired signal spectrum.

In the absence of interference, most signals are smaller than the mean value μ plus twice the standard deviation σ; when there is a series of interference or the total energy of the interference increases, the mean value rises relative to the mean of the Gaussian distribution part, and the standard deviation is also due to increased by the presence of interference. This distortion requires selecting a threshold closer to the mean to filter out the frequency components associated with the interfering signal.

The method of determining the filtering threshold is as follows: first, for each 1024-point FFT data block, calculate the decibel value of the amplitude of each frequency point in the block, and then calculate the standard deviation σ and mean value μ of the block. Select the appropriate weighting factor N from the preset weighting set according to σ, and set the interference detection gate as:

Th=μ+Nσ(3)

In the formula: N is the weighting factor.

The standard deviation is divided into 5 levels σ ₀ , σ ₁ , σ ₂ , σ ₃ , σ ₄ , and N is also divided into N ₀ , N ₁ , N ₂ , N ₃ , N ₄ , and the choice of N is related to σ.

$\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\\ {\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\\ {\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

The preset weighting factor set can be adjusted according to different channel environments, so that after the signal undergoes FFT, N is selected according to the mean value and standard deviation to obtain the interference detection threshold. In the N-sigma algorithm, the determination and selection criteria of N are the key to the algorithm design. If the value of N is too large, the estimation threshold will be too high, causing serious interference leakage; otherwise, if the value of N is too small, the estimation threshold will be too low, resulting in misjudgment.

Larger standard deviations indicate stronger interference, and a small scaling factor is needed to keep the gate on top of and close to the noise. Each frequency point is compared with the threshold one by one, the area greater than the threshold is recorded and its center position is found, and the center position is the center frequency of the normalized interference. Since the values of the five candidate proportional coefficients (N ₀ -N ₄ ) are determined by FPGA programming, the algorithm can be easily applied to other system applications.

2.3 Approximate Calculation of Statistical Engineering of Variance Mean

Interference detection analysis calculates the mean and standard deviation for each FFT data block, and these statistics are used to calculate the threshold. In order to calculate these statistics, the output decibel value of each FFT data block is 10log ₁₀ (|X(k)|), where |X(k)| is the amplitude of the sampled intermediate frequency signal X(k) where logarithmic coordinates are used , which is used to reduce the spectral distortion of the signal mixed by spread spectrum signal, thermal noise and interference, and at the same time reduce the numerical accuracy requirements of |X(k)| required for the subsequent calculation of statistics. In order to reduce the complexity of the hardware implementation, some approximations are made to the computation of 10log ₁₀ (|X(k)|) and statistics. It is verified by simulation that the accuracy drop caused by approximate calculation does not affect the identification of interference signals from spread spectrum signals and thermal noise mixed signals. The approximate calculation of magnitude is as follows:

|X(k)|≈[max(ReX(k), ImX(k))+min(ReX(k), ImX(k))/4]    (5)

Wherein Re is the real part of the signal, Im is the imaginary part of the signal, the overall error of the above approximate operation is 0.6%, and the maximum error is 11.6% (at π/4, 3π/4, 5π/4, 7π/4 respectively).

The second approximation is logarithmic quantization, the formula:

10log ₁₀ (|X(k)|)＝10log ₁₀ (2)·log ₂ (|X(k)|)≈3log ₂ (|X(k)|) (6)

It can be seen that the amplitude calculation can be completed only by shifting and adding operations, while avoiding multiplication and square root operations that are sensitive to hardware resources. The overall error of the above-mentioned logarithmic operations is 0.5%, and the maximum error is 2.9%.

The mean and standard deviation of each FFT data block are calculated as follows:

$\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}{\mathrm{<span\; class="notranslate">\; log</span>}}_{\mathrm{<span\; class="notranslate">\; 10</span>}}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

${\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}}<span\; class="notranslate">\; [</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\mathrm{<span\; class="notranslate">\; 10</span>}}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}}<span\; class="notranslate">\; (</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; P</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\mathrm{<span\; class="notranslate">\; 10</span>}}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

Where N _P is the number of frequency points of each FFT data block. According to the above equation, the accumulator accumulates N _P frequency points 10log ₁₀ (|X(k)|) and (10log ₁₀ (|X(k)|)) ² , after each block is processed, the accumulated value is used to calculate the mean value μ and variance σ ² , and the standard deviation σ can be obtained by the square root of σ ² .

3 Design of interference suppression module based on digital zero-IF

In order to filter out the signal of the interference point, the sampled intermediate frequency input signal needs to be decomposed orthogonally. After the decomposition, the intermediate frequency bandpass signal is moved to the zero intermediate frequency, and the baseband signal is obtained through low-pass filtering. These conversion processes are performed by the digital quadrature in the FPGA. downconverter DDC to complete. The baseband signal is notched at zero frequency through spectrum shifting to filter out strong narrowband interference signals. In order to facilitate subsequent demodulation and despreading processing, the baseband signal needs to be reconstructed into an intermediate frequency bandpass signal (real signal).

3.1 Construct baseband signal

In this design, the system receives an intermediate frequency input signal of 70MHz, and the digital sampling rate is 61.44MHz. The position of the intermediate frequency after digitization by the modulus A/D is F ₀ =8.56MHz, so the center frequency of the band-pass filter is selected as 8.56MHz, and the filter The bandwidth is selected according to different information rates, which is equal to the bandwidth of the narrowband modulation signal without spreading the spectrum. In order to filter out the signal of the interference point, it is necessary to convert the intermediate frequency bandpass signal into a baseband signal, and through DDC quadrature demodulation, the sampling signal is multiplied (mixed) with the quadrature signal output by the on-chip sincos lookup table (SINCOS LUT) , and then pass through a low-pass filter to obtain the baseband signal, as shown in Figure 5.

Let the mathematical expression of the intermediate frequency bandpass signal be:

$\begin{array}{c}\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ <span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>}<span\; class="notranslate">\; +</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>}\end{array}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 9</span>}<span\; class="notranslate">\; )</span>$

where a(t) is the information sequence, f ₀ is the center frequency of the carrier, is the carrier phase, then the frequency shift filter

In the formula S _BI (t) and S _BQ (t) are the co-directional component and quadrature component (I component and Q component) of the intermediate frequency bandpass signal after frequency shift, respectively, is the frequency shift factor. .

The process of shifting the spectrum of the baseband signal constructed from the intermediate frequency bandpass signal is shown in Figure 6. Figure 6a is the spectrum of the sampled intermediate frequency bandpass signal, Figure 6b is the frequency spectrum of the signal after frequency shifting, and Figure 6c is the amplitude-frequency characteristic of the low-pass filter. 6d is the frequency spectrum of the baseband signal generated by the frequency-shifted signal through low-pass filtering.

Among them, the bandwidth of the intermediate frequency band-pass signal is 2B, and f _C is its center frequency.

3.2 Interference point suppression

Assuming that the baseband signal S _B (t) contains the interference component f _I , then the baseband signal and the frequency shift factor After multiplication, the baseband signal S _B ^* (t) with f _I moved to zero frequency is obtained. After S _B ^* (t) is high-pass filtered, the baseband signal S _B '(t) with interference f _I suppressed is obtained. The process is shown in Figure 7. The formula for calculating the baseband signal S _B '(t) that suppresses the interference f _I is as follows:

(S _BI +jS _BQ )×[cos(ω ₀ t)+jsin(ω ₀ t)] (11)

＝[S _BI ×cos(ω ₀ t)-S _BQ ×sin(ω ₀ t)]+j[S _BI ×sin(ω ₀ t)+S _BQ ×cos(ω ₀ t)]

In the formula, ω ₀ =-2πf _I , and the spectra of S _B (t), S _B ^* (t), and S _B '(t) during the interference suppression process are shown in Figure 8.

3.3 IF Bandpass Signal Reconstruction

The signal after interference suppression is a baseband signal (complex signal), which needs to be converted into an intermediate frequency bandpass signal (real signal) for subsequent processing. By analyzing the intermediate frequency bandpass signal S(t), analyzing the signal S _A (t), baseband The relationship between the signals S _B (t) can obtain the mathematical expression of reconstructing the real signal from the baseband signal. The analytical signal S _A (t) can be expressed as:

$\begin{array}{c}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; A</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 12</span>}<span\; class="notranslate">\; )</span>\\ <span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; B</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{\mathrm{<span\; class="notranslate">\; j</span>}\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 13</span>}<span\; class="notranslate">\; )</span>\\ <span\; class="notranslate">\; =</span><span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BI</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BQ</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 14</span>}<span\; class="notranslate">\; )</span>\\ <span\; class="notranslate">\; =</span><span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BI</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BQ</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\\ <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BI</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; BQ</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; cos</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 15</span>}<span\; class="notranslate">\; )</span>\end{array}$

Make the real part of the above formula (12) and (15) equal, that is, the required real signal is:

S(t)＝S _BI (t)cos(2πf _c t)-S _BQ (t)sin(2πf _c t) (16)

in is the Hilbert transform of the intermediate frequency bandpass signal S(t), and _fc is the center frequency of the intermediate frequency bandpass signal.

The process of reconstructing the real signal from the baseband signal is shown in Figure 9. FIG. 10 shows the spectrum of the intermediate frequency bandpass signal S(t), the analysis signal S _A (t), and the baseband signal S _B (t).

The system verification adopts hardware-in-the-loop simulation platform, and the model adopts the digital intermediate frequency part of the digital microwave receiver. The relevant simulation and test parameters are as follows: the data source uses a uniformly distributed random signal, the spreading code is a balanced 1023-bit gold sequence, the code rate is 10.24M, the information modulation method is BPSK, the spreading processing gain is 20dB, and the receiving intermediate frequency The signal is 70MHz, the digital sampling rate is 61.44MHz, the position of the intermediate frequency after A/D digitization is 8.56MHz, the center frequency of the band-pass filter is 8.56MHz, and the filter bandwidth is equal to the bandwidth of the narrow-band modulation signal without spreading .

The simulation test results are as follows: the background noise without IF signal input is shown in Figure 11. In the case of an IF signal input, before notch processing, we can see that there are two obvious NBI peaks in the frequency spectrum of the IF signal, as shown in Figure 12. After IF adaptive notch processing, we can see that NBI is well suppressed on the signal spectrum, as shown in Figure 13.

The present invention focuses on a novel FPGA-based digital zero-IF input signal self-adaptive notch technology, which uses transform domain threshold interference detection technology and digital quadrature demodulation ideas to perform real-time sampling and spectrum analysis on the IF input signal , and preset parameters through the program, select a suitable suppression threshold, move the signal to zero frequency for filtering, and after NBI filtering, restore the baseband signal to an intermediate frequency bandpass signal for demodulation and despreading to realize NBI adaptive to DDDS inhibition. The specific theoretical calculation and derivation process of adaptive notch and related spectrum analysis are given in this paper. The effectiveness of the method for NBI suppression is verified by simulation analysis.

Although the present invention has been particularly shown and described in conjunction with preferred embodiments, it will be understood by those skilled in the art that changes in form and details may be made to the present invention without departing from the spirit and scope of the invention as defined by the appended claims. Making various changes is within the protection scope of the present invention.

Claims (3)

1., based on a FPGA new digital zero intermediate frequency Adaptive notch filtering method, comprise the following steps:

Step 1: utilize FFT to convert and the intermediate-freuqncy signal of sampling is transformed to frequency domain;

Step 2: carry out the analysis of frequency spectrum statistic and threshold judgement to intermediate-freuqncy signal in frequency domain, judges strong arrowband interference NBI number and centre frequency thereof;

Step 3: according to threshold judgement result, if not interference, forwards the original intermediate-freuqncy signal of step 7 to sampling to and directly exports; If there is interference, then carry out NBI AF panel, carry out quadrature demodulation by the intermediate-freuqncy signal of orthogonal digital low-converter to sampling, and utilize low pass filter only to retain the positive frequency component of sampling midband messenger, conversion obtains baseband signal;

Step 4: carry out frequency spectrum shift to baseband signal, moves first noise spot to zero-frequency, by high pass filter filtering interfering spectrum component, thus curbs first noise spot;

Step 5: if noise spot is more than one, repeat the operation of step 4, the signal be about to after first time or the process of second time AF panel carries out frequency spectrum shift again, moves noise spot to zero-frequency, by high pass filter filtering interfering spectrum component, thus curb other noise spots;

Step 6: the signal after AF panel process is transformed to real signal;

Step 7: carry out demodulation output by baseband processing module.

2. according to claim 1 based on FPGA new digital zero intermediate frequency Adaptive notch filtering method, it is characterized in that: the threshold judgement in described step 2 comprises the steps:

Step 21: set up signal model, for direct-sequence communications system, at transmitting terminal by being multiplied with pseudo random sequence by signal, signal extension on a wide frequency spectrum; At receiving terminal, by identical PN sequence is multiplied with Received signal strength, thus restoring signal; Wherein, the intermediate-freuqncy signal x (t) of reception is generally made up of 3 parts:

x(t)＝s(t)+i(t)+n(t) (1)

Wherein s (t) is spread-spectrum signal, and i (t) is arrowband interference, and n (t) is additive white Gaussian noise;

X (t), with after the sampling of PN chip rate, is expressed as:

x(k)＝s(k)+i(k)+n(k) (2)

Wherein s (k) is uncorrelated in time with n (k), and s (k) and i (k) have correlation;

Step 22: FFT data are divided into different data blocks according to every 1024 points, to the FFT data block of each 1024, the decibel value of each Frequency point amplitude in calculated data block, then calculates standard variance σ and the average μ of this data block; Choose suitable weighted factor N according to standard variance σ from the weighted aggregation preset, setting Interference Detection door is:

Th＝μ+Nσ (3)

In formula, N is weighted factor;

σ divides 5 horizontal σ ₀, σ ₁, σ ₂, σ ₃, σ ₄, N is divided into N ₀, N ₁, N ₂, N ₃, N ₄, the selection of N is relevant with σ:

$N=\left\{\begin{array}{c}{N}_0,\mathrm{\&sigma;}<{\mathrm{\&sigma;}}_0\\ N_1,{\mathrm{\&sigma;}}_0<\mathrm{\&sigma;}<{\mathrm{\&sigma;}}_1\\ N_2,{\mathrm{\&sigma;}}_1<\mathrm{\&sigma;}<{\mathrm{\&sigma;}}_2\\ N_3,{\mathrm{\&sigma;}}_2<\mathrm{\&sigma;}<{\mathrm{\&sigma;}}_3\\ N_4,{\mathrm{\&sigma;}}_3<\mathrm{\&sigma;}<{\mathrm{\&sigma;}}_4\end{array}\right.---\left(4\right)$

Step 23: engineering approximation calculating is carried out to the statistic of the standard variance σ in step 22 and average μ: Interference Detection analysis is that each FFT data block calculates standard variance σ and average μ, and it is 10log that each FFT data block exports decibel value ₁₀(| X (k) |); Intermediate-freuqncy signal X (k) amplitude after sampling | X (k) | approximate calculation as follows:

|X(k)|≈[max(ReX(k),ImX(k))+min(ReX(k),ImX(k))/4] (5)

Wherein Re is the real part of signal, and Im is the imaginary part of signal;

The approximate calculation of logarithmic quantization is as follows:

10log ₁₀(|X(k)|)＝10log ₁₀(2)·log ₂(|X(k)|)≈3log ₂(|X(k)|) (6)

The average of each FFT data block and being calculated as follows of standard variance:

$\mathrm{\&mu;}=\underset{k=0}{\overset{N_P-1}{\mathrm{\&Sigma;}}}\frac{\left(10{\log }_{10}\left(\right|X\left(k\right)\left|\right)\right)}{N_P}---\left(7\right)$

${\mathrm{\&sigma;}}^2=\frac{1}{N_P}[\underset{k=0}{\overset{N_P-1}{\mathrm{\&Sigma;}}}{(10\&CenterDot;{\log }_{10}\left(\right|X\left(k\right)\left|\right))}^2-\frac{1}{N_P}\left(\underset{k=0}{\overset{N_P-1}{\mathrm{\&Sigma;}}}{(10\&CenterDot;{\log }_{10}\left(\right|X\left(k\right)\left|\right))}^2\right)]---\left(8\right)$

Wherein, N _pfor the frequency number of each FFT data block; According to above-mentioned equation, accumulator adds up N _pthe 10log of individual frequency ₁₀(| X (k) |) and (10log ₁₀(| X (k) |)) ², each piece process after, accumulated value is used for computation of mean values μ and variances sigma ², standard variance σ is by σ ²evolution obtains.

3. according to claim 1 based on FPGA new digital zero intermediate frequency Adaptive notch filtering method, it is characterized in that: the NBI signal suppressing in described step 3 comprises the following steps:

Step 31, structure baseband signal: make the mathematic(al) representation of midband messenger be:

$\begin{array}{c}S\left(t\right)=a\left(t\right)\cos [2\mathrm{\&pi;}{f}_{0}t+\mathrm{\&phi;}\left(t\right)]\\ =\frac{1}{2}a\left(t\right)e^{j[2\mathrm{\&pi;}{f}_{0}t+\mathrm{\&phi;}\left(t\right)]}+\frac{1}{2}a\left(t\right)e^{-j[2\mathrm{\&pi;}{f}_{0}t+\mathrm{\&phi;}\left(t\right)]}\end{array}---\left(9\right)$

Wherein a (t) is information sequence, f ₀centre carrier frequency, for carrier phase, so filtered baseband signal S of frequency displacement _bt the mathematic(al) representation of () is:

In formula s _bI(t) and S _bQt () is I component and the Q component of former intermediate-freuqncy signal, it is frequency-shifting operator;

Step 32: noise spot is suppressed: by baseband signal S _bt interference component that () comprises is designated as f _i, so baseband signal and frequency-shifting operator f is obtained after being multiplied _imoved to the baseband signal s* (t) of zero-frequency, s* (t) is inhibited interference f after high-pass filtering _ibaseband signal s'(t); Curb interference f _ibaseband signal s'(t) computing formula as follows:

$\begin{array}{c}(S_{\mathrm{BI}}+j{S}_{\mathrm{BQ}})\&times;[\cos \left({\mathrm{\&omega;}}_{0}t\right)+j\sin \left({\mathrm{\&omega;}}_{0}t\right)]\\ =[S_{\mathrm{BI}}\&times;\cos \left({\mathrm{\&omega;}}_{0}t\right)-S_{\mathrm{BQ}}\&times;\sin \left({\mathrm{\&omega;}}_{0}t\right)]+j[S_{\mathrm{BI}}\&times;\sin \left({\mathrm{\&omega;}}_{0}t\right)+S_{\mathrm{BQ}}\&times;\cos \left({\mathrm{\&omega;}}_{0}t\right)]\end{array}---\left(11\right);$

ω in formula ₀=-2 π f _i;

Step 33: midband messenger reconstructs: the signal after AF panel is baseband signal, in order to subsequent treatment need be converted into real signal, by analyzing midband messenger S (t), analytic signal S _a(t), baseband signal S _bt the correlation between (), namely obtains the mathematic(al) representation being reconstructed real signal by baseband signal; Analytic signal S _at () is expressed as:

$\begin{array}{c}{S}_A\left(t\right)=S\left(t\right)+j\hat{S}\left(t\right)\\ =S_B\left(t\right)e^{j2\mathrm{\&pi;}{f}_{c}t}\\ =[S_{\mathrm{BI}}\left(t\right)+j{S}_{\mathrm{BQ}}\left(t\right)][\cos \left(2\mathrm{\&pi;}{f}_{c}t\right)+j\sin \left(2\mathrm{\&pi;}{f}_{c}t\right)]\\ =[S_{\mathrm{BI}}\left(t\right)\cos \left(2\mathrm{\&pi;}{f}_{c}t\right)-S_{\mathrm{BQ}}\left(t\right)\sin \left(2\mathrm{\&pi;}{f}_{c}t\right)]\\ +j[S_{\mathrm{BI}}\left(t\right)\sin \left(2\mathrm{\&pi;}{f}_{c}t\right)+S_{\mathrm{BQ}}\left(t\right)\cos \left(2\mathrm{\&pi;}{f}_{c}t\right)]\end{array}---\left(12\right)$

Wherein the Hilbert transform of midband messenger S (t), f _cfor the centre frequency of midband messenger;

Relatively (12) formula equation two ends, obtaining required real signal is:

S(t)＝S _BI(t)cos(2πf _ct)-S _BQ(t)sin(2πf _ct)。