WO2015172622A1

WO2015172622A1 - Method for radio-frequency interference suppression of high-frequency ground wave radar

Info

Publication number: WO2015172622A1
Application number: PCT/CN2015/077174
Authority: WO
Inventors: 陈泽宗; 谢飞; 易盛; 赵晨; 曾耿斐; 王方俊
Original assignee: 武汉大学
Priority date: 2014-05-14
Filing date: 2015-04-22
Publication date: 2015-11-19
Also published as: AU2015101677A4; CN103954944A

Abstract

Disclosed is a method for radio-frequency interference suppression of high-frequency ground wave radar. On the basis of original high-frequency ground wave radar signal processing, an empirical mode decomposition method is utilized to decompose a received signal into a plurality of intrinsic mode functions, radio-frequency interference positions are detected through a standard deviation and sliding window processing, sampling points at the radio-frequency interference positions of the intrinsic mode functions are subjected to power spectrum and instantaneous frequency determination, and the amplitude of the sampling points meeting the requirements is set to be zero. The method has outstanding application effect in actual data processing, causes no damage to signals during noise reduction, brings a significant suppression effect, increases the range of radar detection, and raises the precision of radar detection.

Description

Method for suppressing radio frequency interference of high frequency ground wave radar

Technical field

The invention belongs to the field of radar signal processing, and particularly relates to a method for suppressing radio frequency interference of high frequency ground wave radar.

Background technique

HF ground wave radar is an important detection equipment for continuous large-area marine environment monitoring. It mainly detects ocean dynamic parameters such as wind, waves, current and tide on the ocean surface and low-altitude low-speed moving targets on the sea surface. 32,000 kilometers, so high-frequency ground wave radar is of great significance to the national economy and national defense construction.

HF ground wave radar generally works at 3-30MHz, and there will be a lot of radio frequency interference in this frequency band. The radio frequency interference mainly comes from dense short-wave communication signals, broadcast station signals and industrial interference in the high frequency band of radar operation. Since these disturbances are all active, their power is very powerful compared to the radar echo signal. When it enters the receiver, it will greatly reduce the data quality of the high-frequency radar, which will seriously affect the subsequent extraction of ocean surface dynamic parameters. Obstruction, sometimes even unable to extract information such as wind, waves, and flow.

Radio frequency interference mainly has the following characteristics: (1) In the distance Doppler spectrum, it appears as a longitudinal strip distributed along the distance axis, which is basically reflected in all distance elements. Random information components in radio frequency interference contribute to the noise floor. (2) The time when radio frequency interference is more serious generally occurs between 17:00 and 21:00 in the evening. At this time, due to the disappearance of the D layer of the ionosphere, the communication signals of short-wave radio stations, especially busy fishing vessels, cannot be effectively shielded by the ionosphere. The radar system interferes with the normal operation of the radar. Since the short-wave frequency resources are very limited, it is very difficult for the radar to select a frequency band that does not contain interference, so it is particularly important to remove or suppress the radio frequency interference in the echo.

Among the existing high-frequency radar anti-radio frequency interference methods, there are mainly adaptive methods and time-frequency domain anti-interference methods. In the adaptive beamforming anti-interference method, the system is relatively complicated because it relies on a large phased antenna array, and is not suitable for a small-caliber wide beam radar. Therefore, this method is rarely used. Anti-interference methods based on signal time-frequency domain characteristics, such as detection based on transient interference-removal-recovery method and interference-space sub-space orthogonal projection method based on interference distance distribution characteristics, all require reception of radio frequency interference The signal is filtered, which inevitably causes a certain degree of loss and distortion to the useful signal (such as increased sidelobe amplitude, main lobe broadening), and this effect becomes more serious when there are multiple disturbances. The suppression effect is not good.

Summary of the invention

The present invention is directed to the above problem, and proposes a method for suppressing radio frequency interference of a high frequency ground wave radar, which avoids the adverse effect of the existing anti-radio frequency interference algorithm on the loss of the radar signal. The invention completely preserves the useful signal while reducing the interference, and provides a fast and effective radio frequency anti-interference method for the existing high-frequency radar system, thereby increasing the radar detection range and improving the radar detection precision.

The technical solution of the present invention is a method for suppressing radio frequency interference of a high frequency ground wave radar, comprising the following steps:

Step 1, setting a frequency threshold f _max according to the radar waveform parameter and the maximum detection distance of the radar;

Step 2, performing empirical mode decomposition EMD (Empirical Mode Decomposition) on the received original signal to obtain two or more eigenmode functions IMF (Intrinsic Mode Function) and a trend function r _n (t);

Step 3: The first eigenmode function IMF, which is most affected by radio frequency interference, detects radio frequency interference by standard deviation determination and sliding window processing, and marks the location of radio frequency interference;

Step 4: Select an eigenmode function IMF, perform a Hilbert transform, and find the instantaneous frequency corresponding to each time point of the eigenmode function IMF, namely:

[A f]=hilbert(IMF)

Step 5: Find the average amplitude A _{mean of} the eigenmode function IMF in step 4, and set the amplitude of the interference position point to zero when the amplitude of the interference position point is higher than the triple average amplitude A _mean , that is, IMF (A>3*A _mean )=0; otherwise, if the instantaneous frequency of the interference position point is higher than the set frequency threshold f _max , the amplitude of the interference position point is also set to zero, that is, IMF (f ≥f _max )=0; after traversing the sampling points of all the interference positions, go to step 6;

Step 6, repeating

steps

4 and 5 for all eigenmode functions IMF until the processing of all eigenmode functions IMF is completed, go to step 7;

In step 7, all the processed IMFs and the original trend items (trend functions) are added to obtain a signal after interference suppression.

In the step 3, the standard deviation σ of the first eigenmode function IMF is obtained. In the first eigenmode function IMF, whether the eigenmode function IMF sliding window is used to determine whether it is higher than The standard deviation σ is used to determine the location of the RF interference.

In the step 3, the length of the window is set to L, and the standard deviation of L points in the window is obtained.

If the standard deviation of one of the points

Above σ, it is determined that this is radio frequency interference, otherwise, the next sliding window is continued, thereby detecting the location of the interference.

The advantages of the invention are:

1. In the high-frequency ground wave radar, the EMD algorithm is used for de-interference processing. The EMD is used to decompose the received signals into different IMFs according to the frequency, and each IMF is processed to suppress the radio frequency interference while maximally Useful signal characteristics are preserved.

2. For the first IMF processing, the position of the interference is detected by standard deviation and sliding window processing, and the position of the radio frequency interference can be accurately detected.

3. The standard deviation and the instantaneous frequency are determined for the sampling points at the interference position of all eigenmode functions IMF, which reduces false positives and misses, and maximizes the interference suppression effect.

4. Zero the amplitude of the sampling point at the interference position and satisfying the above conditions, without losing the useful signal, and the speed is fast, meeting the requirements of real-time operation of the radar.

The invention proposes a novel radio frequency interference suppression method for high-frequency ground wave radar through the above innovations. This method should have a good effect in actual data processing, and the noise is reduced while the signal is suppressed, and the suppression effect is remarkable, and the method is increased. Radar detection distance improves radar detection accuracy.

DRAWINGS

Figure 1. Block diagram of the working principle of HF ground wave radar;

Figure 2, the process of radio frequency interference processed by the system;

Wherein, B-sweep bandwidth, T-sweep period, t ₀ - interference duration, f _rfi - interference frequency, f ₀ - operating frequency, b-filter bandwidth;

Figure 3 is a flow chart of the algorithm of the present invention;

Figure 4, the distance spectrum of (severe) radio frequency interference in actual reception;

Figure 5. Distance spectrum after RF interference suppression.

detailed description

The present invention will be further described below by way of specific examples in conjunction with the accompanying drawings.

The working principle of high frequency ground wave radar and the characteristics of radio frequency interference in linear frequency modulation system are introduced as follows:

The working principle block diagram of HF ground wave radar is shown in Figure 1 (1 is the receiving antenna, 2 is the local oscillator signal of linear sweep, 3 is low-pass filtering, 4 is solution distance transform, 5 is Doppler transform) . Radar line In the frequency modulation waveform system, the echo spectrum is demodulated, low-pass filtered, sampled, and fast Fourier transform in the fast time domain in each sweep period to obtain the distance spectrum of the sweep period. Each spectral point in the distance spectrum corresponds to a sampling point of a distance element, and the discrete Fourier transform of the sampling sequence of the plurality of distance spectra is performed in a coherent accumulation time of the plurality of sweeping periods to obtain the distance. The Doppler spectrum of the Yuan. A distance-Doppler two-dimensional spectrum is obtained by performing a discrete Fourier transform of the slow time domain for all distance elements in the detection range.

The chirp local oscillator signal can be expressed as

Where f ₀ is the radar carrier frequency, K=B/T is the sweep frequency, B is the sweep bandwidth, and T is the sweep width.

Without loss of generality, consider radio frequency interference as a single frequency signal,

The interference signal output after mixing and low-pass filtering is

It can be seen from the above equation that the demodulated and filtered output of a single-frequency interference is a bandwidth-limited chirp signal. The bandwidth of this signal is the bandwidth b of the low pass filter.

Figure 2 illustrates the process by which the RF interference at frequency f _rfi is processed by the system. The radio frequency interference in the same frequency band as the radar operating frequency is mixed with the chirped local oscillator signal to generate a new chirp signal, which is obviously short-lived in the time domain after passing through the low-pass filter, and its duration

The base noise is greatly increased in the frequency domain (i.e., the distance spectrum), and the interference appears as an interference band parallel to the distance axis on the distance Doppler spectrum.

If there are multiple RF interferences as follows:

Where f _c is the carrier, f _i is the frequency deviation of a certain frequency component relative to the carrier, and a _i is the amplitude of the ith frequency component,

For the initial phase, N is the number of different frequency interferences.

A plurality of interference bands parallel to the distance axis will appear on the Doppler spectrum, which will usually flood the Bragg peak, thus seriously affecting the radar detection performance.

The invention is characterized by short time and high intensity after demodulation by radio frequency interference, and time series of echoes The empirical mode is decomposed into multiple IMFs, and interference suppression is performed separately, and the signal characteristics are retained to the greatest extent. See Figure 3 for specific implementation steps:

Step 2: performing empirical mode decomposition (EMD) on the received original signal to obtain n (n≥2) eigenmode functions (IMF) and a trend function r _n (t);

Steps of the EMD decomposition method:

1 The maximum value and the minimum value of the signal x(t) are respectively connected by a cubic spline function to form an upper and a lower envelope;

2 For each moment, take the average of the upper and lower envelopes to form the signal m ₁ (t), then subtract m ₁ (t) from the signal x(t) to form another signal h ₁ (t), ie:

x(t)-m ₁ (t)=h ₁ (t) (5)

3 pairs of signals h ₁ (t)

repeat

1, 2 get the signal h ₂ (t), namely:

h ₁ (t)-m ₂ (t)=h ₂ (t) (6)

Here, m ₂ (t) is a signal formed by the average value of the upper and lower envelopes of h ₁ (t).

4 Repeat steps 1, 2, and 3 above until the signal h _k (t) obtained (for example, after k repetitions) is an IMF, namely:

h _k (t)=h _k-1 (t)-m _k (t) (7)

Thus, the first modality h _k (t) is decomposed from the signal x(t) and recorded as C ₁ (t), and such a process is called a screening process;

5 Subtracting x(t) from the first modality C ₁ (t) to form a new signal r ₁ (t), namely:

x(t)-C ₁ (t)=r ₁ (t) (8)

Repeating steps

1, 2, 3, and 4 for r ₁ (t) to calculate a second mode C ₂ (t);

6 Subtracting r ₁ (t) from C ₂ (t) forms a new signal r ₂ (t), namely:

r ₁ (t)-C ₂ (t)=r ₂ (t) (9)

Repeating steps

1, 2, 3, and 4 for r ₂ (t) to calculate a third mode C ₃ (t);

7 Repeat steps 1, 2, 3, and 4 until the last r _n (t) is a monotonic function, thus decomposing x(t) out of n IMFs and a trend function r _n (t).

Step 3, set the window function:

Where N is the data length of the IMF and L is the length of the window function, which is set to 4. Find the standard deviation σ of the first IMF, and set σ as the detection threshold of transient interference. The rectangular window then slides over the first IMF to calculate the standard deviation of the data in each sliding window.

when

It is determined that the echo data in the segment contains interference, and if not, the next sliding window is continued, thereby detecting the location of the interference.

Step 4: Perform a Hilbert transform on any IMF to find the instantaneous frequency corresponding to each time point of the IMF, namely:

[A f]=hilbert(IMF) (11)

Where A represents the amplitude corresponding to each time point and f represents the instantaneous frequency corresponding to each time point.

In step 5, the average amplitude A _{mean of} the IMF is obtained. When the interference position is higher than the triple amplitude A _mean , the amplitude of the point is set to zero, that is, IMF (A>3*A _mean )=0. Otherwise, if the instantaneous frequency of the interference location point is higher than the set frequency threshold f _max , the amplitude of the point is also set to zero, ie IMF (f ≥ f _max ) = 0; traversing the sampling points of all interference locations After that, go to step 6;

Step 6. Repeat steps 4 and 5 for all IMFs. Until the processing of all IMFs is completed, go to step 7;

In step 7, all the processed IMF components and the original trend items are added to obtain a signal after interference suppression.

Figure 4 shows the distance spectrum of radio frequency interference. Figure 5 shows the distance spectrum after radio frequency interference suppression. It can be seen that the radio frequency interference is greatly suppressed, and the signal is not lost, and the signal-to-noise ratio is significantly improved.

Claims

A method for suppressing radio frequency interference of a high frequency ground wave radar, comprising: the following steps:

Step 1, setting a frequency threshold f max according to the radar waveform parameter and the maximum detection distance of the radar;

Step 2, performing empirical mode decomposition EMD on the received original signal to obtain two or more eigenmode functions IMF and a trend function r n (t);

Step 3: The first eigenmode function IMF, which is most affected by radio frequency interference, detects radio frequency interference by standard deviation determination and sliding window processing, and marks the location of radio frequency interference;

Step 4: Select an eigenmode function IMF, perform a Hilbert transform, and find the instantaneous frequency corresponding to each time point of the eigenmode function IMF, namely:

[A f]=hilbert(IMF)

Step 5: Find the average amplitude A mean of the eigenmode function IMF in step 4, and set the amplitude of the interference position point to zero when the amplitude of the interference position point is higher than the triple average amplitude A mean , that is, IMF (A>3*A mean )=0; otherwise, if the instantaneous frequency of the interference position point is higher than the set frequency threshold f max , the amplitude of the interference position point is also set to zero, that is, IMF (f ≥f max )=0; after traversing the sampling points of all the interference positions, go to step 6;

Step 6, repeat steps 4 and 5 for all eigenmode functions IMF until the processing of all eigenmode functions IMF is completed, go to step 7;

In step 7, all the processed IMFs and the original trend function are added to obtain a signal after interference suppression.
A method for radio frequency interference suppression of a high frequency ground wave radar according to claim 1, wherein in step 3, a standard deviation σ of the first eigenmode function IMF is obtained, in the first In the eigenmode function IMF, it is determined whether the radio frequency interference position is determined by determining whether the eigenmode function IMF sliding window is higher than the standard deviation σ.
The method for suppressing radio frequency interference of a high-frequency ground wave radar according to claim 1 or 2, wherein in the step 3, the length of the window is set to L, and the standard deviation of L points in the window is obtained.
If the standard deviation of one of the points
Above σ, it is determined that this is radio frequency interference, otherwise, the next sliding window is continued, thereby detecting the location of the interference.
A method for suppressing radio frequency interference of a high frequency ground wave radar according to claim 1, wherein

The steps of the empirical mode decomposition EMD in the step 2 are as follows.

1 The maximum value and the minimum value of the received original signal x(t) are respectively connected by a cubic spline function to form an upper and a lower envelope;

2 For each moment, take the average of the upper and lower envelopes to form the signal m 1 (t), then subtract m 1 (t) from the signal x(t) to form another signal h 1 (t), ie:

x(t)-m 1 (t)=h 1 (t)

3 pairs of signals h 1 (t) repeat 1, 2 get the signal h 2 (t), namely:

h 1 (t)-m 2 (t)=h 2 (t)

Where m 2 (t) is a signal formed by the average of the upper and lower envelopes of h 1 (t);

4 Repeat steps 1, 2, and 3 above until the obtained signal h k (t) is an IMF, that is:

h k (t)=h k-1 (t)-m k (t)

Thus, the first modality h k (t) is decomposed from the signal x(t) and recorded as C 1 (t), and such a process is called a screening process;

5 Subtracting x(t) from the first modality C 1 (t) to form a new signal r 1 (t), namely:

x(t)-C 1 (t)=r 1 (t)

Repeating steps 1, 2, 3, and 4 for r 1 (t) to calculate a second mode C 2 (t);

6 Subtracting r 1 (t) from C 2 (t) forms a new signal r 2 (t), namely:

r 1 (t)-C 2 (t)=r 2 (t)

Repeating steps 1, 2, 3, and 4 for r 2 (t) to calculate a third mode C 3 (t);

7 Repeat steps 1, 2, 3, and 4 until the final r n (t) is a monotonic function, thus decomposing x(t) out of more than two IMFs and a trend function r n (t).