Method for eliminating interference signals from received signals of a receiving system
By determining interference spectral lines using sensors and applying Fourier transformation with a time-limited window, the method effectively removes interference in receiving systems without losing useful signal information, addressing the complexity and accuracy issues of existing methods.
Patent Information
- Application Number
- DE4104699
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 1991-02-15
- Publication Date
- 2024-12-05
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing methods for eliminating interference spectral lines in receiving systems, such as sonar and airborne sound systems, require complex processing channels and precise knowledge of phase and amplitude, leading to high computational effort and potential errors in location tasks, especially when interference levels are higher than useful signals.
Determine interference spectral lines using sensors near the interference source, apply Fourier transformation with a time-limited window and frequency interval, and adapt the complex spectrum by setting or leveling coefficients to eliminate interference without losing phase relationships, allowing for digital signal processing in the frequency domain.
Enables efficient interference removal without losing useful signal information, preserving phase relationships, reducing computational complexity, and maintaining accurate detection and bearing capabilities, even with high interference levels.
Smart Images

Figure 00000008_0000 
Figure 00000008_0001 
Figure 00000008_0002
Abstract
Description
[0001] The invention relates to a method for eliminating interference signals from received signals of a receiving system of the type mentioned in the preamble. Waves are emitted by a transmitter to be located and received by a receiving system as a useful signal. The direction of incidence of the waves indicates the bearing to the transmitter. If jammers also emit waves in the same frequency range at the same time, bearing determination becomes difficult or even impossible because both useful and interference signals are received simultaneously. Such receiving systems are used, for example, in underwater sound technology as sonar systems on board a watercraft, e.g. submarines, ships or torpedoes, to locate other watercraft based on their sound radiation, whereas an airborne sound direction finding system is used to detect and locate helicopters, e.g. from a tank. The received signals often consist of useful and interference signals with a pronounced line spectrum.The lines of the useful and jamming signals lie in the same frequency range and are often close together. Particularly disruptive to the evaluation of received signals are the interference lines caused by the carrier vehicle acting as a jammer. These lines are caused by the vehicle's propulsion and are received by the sonar or airborne sound system via the sound transmission medium or the vehicle's walls. The reason for this is that both the useful signals received for detection, direction finding, and location, as well as the jamming signals, exhibit a pronounced line spectrum, and it is precisely the spectral lines in the low-frequency range that are transmitted and received particularly well by the carrier vehicle and the target. Separating the spectral lines into useful and jamming components and eliminating the jamming spectral lines is therefore urgently necessary to meet the detection requirements.
[0002] However, these interference spectral lines must also be eliminated if the useful signal is received and evaluated over a broadband frequency, as is the case with passive sonar systems in the higher frequency range, for example, since their levels are often significantly higher than those of the useful signal.
[0003] It is known from US Pat. Nos. 3,599,142 and 4,232,381 to receive the self-generated noise of the host vehicle as an interference source with a sensor, for example an accelerometer, determine its fundamental lines and harmonics according to magnitude and phase, and subtract them from the received signal. This method is adaptive and adjusts to the respective interference noise, so that it can be used even at different speeds of the moving host vehicle. To determine the frequencies of the interference spectral lines, a phase-lock loop circuit is used, and the output signal is evaluated for amplitude and phase in order to subtract the interference spectral lines from the received signal not only in terms of their magnitude but also in terms of their phase, thus preventing corruption of the desired signal.The effort is considerable, as a separate processing channel must be provided for each interference spectral line. This requires not only complex components and synchronization tests, but also a robust design. The effort increases dramatically if, for example, one such processing channel is to be used for each frequency for each hydrophone on a linear, flat, or cylindrical base of a waterborne sound receiving system.
[0004] US Pat. No. 4,594,695 discloses a receiving system in which interference spectral lines in the received signal are determined by correlating the frequency spectra of the received signal and the interference signal. The interference spectral lines are subtracted from the frequency spectrum of the received signal. The inverse Fourier transform produces the interference-free received signal. Precise knowledge of the phase position of the interference spectral lines in the received signal is essential for this interference removal, since phase errors directly lead to errors in the location of the receiving system.
[0005] It is an object of the present invention to provide a method of the type mentioned in the preamble of claim 1, in which interference spectral lines are adaptively eliminated without loss of information in the useful signal and without knowledge of the amplitude and phase position of the interference spectral lines in the received signal, but only by measuring the frequency.
[0006] This object is achieved according to the invention by the features mentioned in claim 1.
[0007] The frequencies of the interference spectral lines are determined using sensors at the source of the interference. For example, a tachometer or an accelerometer is located near the drive units that are the sources of interference. A subsequent frequency analysis provides the frequencies. In underwater acoustic technology, for example, the frequencies of the interference spectral lines caused by operating units, such as fans, are already known from previous measurements on board the carrier vehicle. (followed by pages 4 to 19 of the original description)
[0008] In this case, a special frequency determination during operation of the sonar system can be dispensed with and this prior knowledge can be used.
[0009] To filter out the lowest-frequency interference spectral line from the received signal, a frequency interval is determined. This frequency interval is a fraction of this lowest frequency and forms the basis for the frequency transformation of the received signal into a complex spectrum. The temporal received signal is transformed into the frequency domain using Fourier transformation. Its frequency resolution is determined by the frequency interval. For the Fourier transformation, a time-limited section of the received signal within a time window is used. The time window is equal to the reciprocal of the frequency interval and thus a multiple of the period of the lowest-frequency interference spectral line. The section is formed by multiplying the received signal by a rectangular function whose length defines the time window.Multiplying the time profile of the received signal by the time window during a transformation from the time domain to the frequency domain corresponds to convolving the complex frequency spectrum of the received signal with an Si function whose variable is the frequency interval and thus depends on the frequency of the interference spectral line. Convolving a spectrum in sampled values with an Si function corresponds to interpolating its sampled values.
[0010] A particular advantage of this signal processing is that the energy of the interference spectral line is concentrated in the Fourier transform coefficient corresponding to this lowest frequency and is not smeared into sidebands. This is because the temporal oscillation corresponding to this interference spectral line fits completely within the time window, and the beginning and end of this oscillation can be joined together without any interruptions. Their instantaneous values at the beginning and end of the time window are therefore equal. The same applies to multiples of this frequency, i.e., to the harmonics. This enforces a deterministic signal with respect to this interference spectral line in the time domain, so that a Fourier transform can be performed error-free, since only a frequency-limited, not also time-limited, signal is transformed with respect to the interference signal at this frequency and its harmonics.According to the invention, this interference spectral line and its harmonics are adjusted to the shape of the complex spectrum, i.e., set to zero if there is no energy at neighboring frequencies, or leveled off according to the level curve. If only a single interference spectral line is present, the spectrum of the received signal is now free of interference, and a back transformation to the time domain yields the useful signal without any loss of information.
[0011] It is particularly advantageous if, according to claim 2, the frequency interval is determined by dividing the frequency of the interference spectral line by an integer that is close to the frequency itself. For example, if the frequency is 25.25 Hz, the frequency interval is Δf=1.01 Hz=25.25:25. The normally present rasterization of the complex spectrum in frequency intervals of 1 Hz shifts to a rasterization of Δf=1.01 Hz.
[0012] After the Fourier transformation, the coefficients indicate the frequency-divided energy of the received signal at the corresponding frequencies. In digital signal processing, the frequency range is divided into a number of frequency cells for equal frequency intervals, whereby, according to the invention, the frequency interval is derived from the first interference spectral line. The number of frequency cells multiplied by the frequency interval indicates the frequency range of the Fourier transformation, which must be at least as large as the frequency range of the received signal. This frequency range simultaneously defines a minimum sampling frequency necessary for digital processing. The sampling frequency is an integer multiple of the frequency interval and at least twice as large as the frequency range of the received signal and thus of the Fourier transformation. The complex spectrum of the received signal is then periodic with respect to the sampling frequency.A particular advantage of the further development according to claim 3 is that the frequency resolution and the sampling frequency are both determined by the frequency of the interference spectral lines, so that the coefficient of the Fourier transformation, which indicates the energy of the interference spectral line at this frequency, can be assigned to the content of a single frequency cell.
[0013] To flatten the interference spectral line and its harmonics, the contents of the frequency cells in the range of the interference spectral line and its harmonics are compared. Their contents are set to zero or, if they are significantly larger than the contents of neighboring frequency cells and stand out clearly from the spectrum, reduced. This allows the interference spectral lines, even if they are close to a useful spectral line, to be detected and completely suppressed without the useful spectral line also disappearing.
[0014] The advantage of the signal processing according to the invention lies in the fact that all process steps are performed in the frequency domain. Phase relationships of the spectral components in the received signal are not lost, since the complex spectrum reproduces the received signal with phase and amplitude fidelity. Suppression of the interference spectral line does not affect neighboring useful spectral lines, since the selection of the frequency interval, time window, and sampling frequency does not cause sidebands related to the frequency of the interference spectral line. Knowledge of the phase relationships between the interference and useful components in the received signal is not necessary; the phase relationships are not influenced or disturbed by the interference suppression, so that the evaluation of the useful signals remains unimpaired.
[0015] According to the advantageous further development according to claim 3, the content of the frequency cell with the interference spectral line is not set equal to zero, but is fitted into the course of the spectrum by averaging, wherein the real and imaginary parts of the spectral line are advantageously processed separately.
[0016] The advantage of the further development according to claim 3 is that, by adapting the coefficients to the spectrum profile, no additional interference occurs during the evaluation of the received signals, but rather, reconstruction is possible without any loss of information. If, for example, the cleaned spectrum is retransformed to the time domain, edge effects in the temporal progression of the cleaned received signal at the beginning and end of the time window defined by the frequency interval are so small that detection and direction finding are not impaired, and the full information of the useful signal in the received signal can be used for evaluation.
[0017] An iterative method for eliminating interference spectral lines, which are distributed, for example, across the entire frequency range, is provided by the advantageous developments of the method according to claims 4 and 5. The interference spectral lines in the complex spectrum of the received signals are successively leveled. First, the frequency interval, the time window, and a first sampling frequency are derived from the lowest frequency in the interference spectrum, and a first spectrum is formed that is cleaned of the associated interference spectral line and its harmonics. This cleaned spectrum is then sampled at a second sampling frequency that is a multiple of the frequency interval derived from the spectral line of the next higher frequency. This new frequency interval determines the frequency resolution of the spectrum obtained from the first cleaned spectrum, rasterized at the distance of the new frequency interval, and then sampled again.The new frequency interval must be assigned to each frequency cell. With the same number of frequency cells, the frequency range of the Fourier transform has changed by a factor resulting from the ratio of the frequency intervals. The sampling frequency has also shifted by this factor. The result is again a periodic line spectrum, with the contents of the frequency cells corresponding to the frequency of the interference spectral line to be eliminated and its harmonics being leveled according to the contents of the surrounding frequency cells. The second, cleaned line spectrum now only contains spectral lines attributable to the wanted signal and higher-frequency interference spectral lines. Interference spectral lines with the lowest and next highest frequencies and their harmonics have already been eliminated.If an interference spectral line still exists at a higher third frequency, the second cleaned spectrum is interpolated and sampled at a third sampling frequency according to the frequency interval derived from the next higher frequency. The resulting spectrum is leveled with respect to this interference spectral line. The order of suppression of the interference spectral lines described here is not mandatory; it can also be performed depending on the level of the interference spectral lines, so that the interference spectral line with the highest level is suppressed first, followed by the one with the lower level.
[0018] The interpolation of each cleaned spectrum in the time domain corresponds to a multiplication of its inverse transform by a time window that has a multiple of the period of the frequency of the interference spectral line to be eliminated next. Convolution with the interpolation function in the frequency domain is computationally much simpler than an inverse transformation and a multiplication in the time domain by a time window, since, due to the existing frequency and time limitations of the received signals, disturbing edge effects would be unavoidable during the inverse transformation. The particular advantage of the embodiment according to claim 6 is that the entire signal processing is performed in the frequency domain, so that these edge effects are not significant. If direction finding and detection are also performed in the frequency domain, an inverse transformation of the interference-free spectrum can be completely omitted.A further advantage of the inventive development of the method according to claim 5 is that interference spectral lines can be eliminated independently of their frequency position within the frequency range and that the frequency interval and the associated sampling frequency lead to orders of magnitude that can be easily handled on a computer.
[0019] The current state of computer technology offers the advantage of implementing the inventive development according to claim 5 simultaneously for the received signals of each individual transducer of an antenna configuration—in underwater acoustic technology, for each hydrophone of a receiving arrangement before directional determination—because Fourier transformation, interpolation, spectrum shift keying, and coefficient blanking are common routines. Synchronization and the associated testing problems that occur in analog technology are not a factor.
[0020] Particularly advantageously, the interference elimination according to the invention can also be carried out after the direction formation if the signal processing is carried out in the frequency domain, as described, for example, in DE-PS 30 17 797.
[0021] In the further development of the inventive method according to claim 6, it is advantageous that the interpolation and sampling are performed in a single processing step by convolving the last cleaned spectrum with a sampled interpolation function. The interpolation function itself has as a variable the ratio of the newly determined and last determined frequency interval and is sampled at a higher sampling frequency than its own periodicity. The sampling frequency is derived from the newly determined frequency interval and is thus dependent on the frequency of the next interference spectral line to be eliminated.
[0022] As is generally known in the literature, triangular functions or, particularly advantageously, an Si function according to claim 7 are suitable for interpolation, since the convolution of the spectrum with an Si function performed in the frequency domain corresponds to a multiplication of the time signal with a rectangular window performed in the time domain. The length of the time window corresponds to the distance between its zero crossings, i.e., to the newly determined frequency interval and thus to the frequency of the interference spectral line to be eliminated.
[0023] According to the refinement of the inventive method according to claim 8, it is particularly advantageous to multiply the Si function by a shift factor, thereby shifting a time window in the time domain. The shift factor is a complex exponential function and has as a variable the quotient of the newly and most recently determined frequency interval. It is set, for example, such that the corresponding time window captures exactly the same-direction zero crossings at the beginning and end of the oscillation to be eliminated.
[0024] A back-transformation of the spectrum, cleansed of all interference spectral lines, into the time domain according to claim 9 results in a received signal that contains only the desired desired signal. A window function similar to a Hanning or Butterworth window eliminates any remaining edge effects of the time signal. The associated spectrum then only exhibits lines at the frequencies of the desired signal.
[0025] The method according to the invention is also suitable for determining frequencies of the interference spectral lines from the received signal, which then only contains the interference signal, at times when no useful signals are being received. An arbitrary frequency interval is specified, thereby generating a first spectrum of the received signal. A first frequency at the highest spectral level and the level value are recorded. New frequency intervals are formed in predeterminable steps, and the first spectrum is thus interpolated and sampled. The spectrum with the relatively highest signal level in the vicinity of this first frequency precisely indicates the first frequency. The corresponding interference spectral line is flattened. The thus cleaned spectrum is then examined for the next interference spectral line, and so on.This method of determining the frequencies of the interference spectral lines is advantageous, for example, in a receiving system that uses a vehicle with running engines as the carrier to track other vehicles. For this measurement, the direction finding system is controlled so that the receiving system is directed toward the carrier vehicle. Thus, it only receives interference signals during this time, especially if the measurement is not performed at the site of use.
[0026] The method according to the invention can be advantageously implemented for a receiving system for detecting and / or locating targets using devices according to claims 0 to 3. Due to the complex digital signal processing, phase relationships of the received signals are preserved, and the computational effort is so low that interference can be eliminated from the individual received signals of the transducers in a multi-channel, time-parallel manner, even before direction determination, so that the false alarm rate and direction finding errors are very low. Even spectral lines in the spectrum of the received signal that are closely adjacent in frequency can be distinguished and separated, thus enabling interference elimination. The accuracy of the signal processing depends primarily on the accuracy of the frequency determination of the interference spectral lines. Interference elimination is possible through prediction or a simple rotational speed measurement.However, there are now significantly more precise frequency measurement methods for adaptive noise cancellation, for example, using parametric spectral analysis. The processing steps used are described, for example, in the book "Digital Spectral Analysis," by S. Lawrence Marple Jr., Prentiss Hall, Ing., New Jersey 1987.
[0027] The method and devices according to the invention are characterized by the fact that a determination of the phase position and / or amplitude values of the interference spectral lines is unnecessary, since interference cancellation is carried out solely on the basis of precise knowledge of the individual frequencies of the interference spectral lines. Thus, knowledge of the propagation of the interference signals from the interference source to the receiving system is superfluous. This means that, in the case of an acoustically operating receiving system, it is not necessary to consider the transmission behavior for the interference noise coupled in as structure-borne noise. A further advantage is that interference cancellation is completely level-independent, and thus the interference energy may be orders of magnitude higher than the useful energy to be evaluated, as is the case, for example, when locating helicopters with an acoustic antenna from a carrier vehicle with running engines.The method according to the invention can also be advantageously applied when the frequency of the interference spectral lines is not constant, since the interference signal spectrum can be re-measured for each time window, which then serves as the basis for eliminating interference from the stored received signal in the same time window. The method can thus be adapted to the respective operating state. The method according to the invention also has the advantage that frequency ratios in the spectrum of the wanted signal are fully preserved, so that a classification of the located target can also be performed using the interference-free spectrum.
[0028] The method according to the invention can also be used in radio reconnaissance to suppress jamming transmitters operating on a frequency close to the desired transmitter. Another application is the elimination of interference from a received signal that is superimposed with hum or switching frequencies.
[0029] The invention is illustrated in more detail using an exemplary embodiment for eliminating interference signals from received signals of a receiving system. The figures show: Fig. 1 a received signal over time, Fig. 2 to 5 magnitude spectra over frequency, Fig. 6 a block diagram.
[0030] Fig. Figure 1 shows the received signal of a receiver transducer of an airborne sound direction finding system over time. The received signal contains useful signals originating from the radiated noise of a helicopter and interference signals generated by the carrier vehicle of the airborne sound direction finding system.
[0031] Fig. Figure 2 shows the magnitude of the spectrum of the received signal, which has a first interference spectral line and its harmonics at the frequencies f1, 2f1, and 3f1, and two further interference spectral lines at the frequencies f2 and f3. The frequencies f1, f2, and f3 are determined from the signal of an interference signal sensor and are extracted from the spectrum by the method according to the invention. Fig. 2 eliminated. The spectral lines of the wanted signal at frequencies of 10 Hz, 20 Hz, 30 Hz, and 45 Hz cannot be measured separately and are not significantly visible in the spectrum. The goal is to ensure that the wanted spectral lines can be clearly identified from the remaining spectrum at the end of the interference cancellation procedure.
[0032] Fig. Figure 6 shows a block diagram used to perform the interference cancellation. A receiving transducer 10, which supplies the received signal, is connected to a Fourier transform circuit 13 via a time gate 11 and a digital-to-analog converter 12. An interference signal sensor 20, located near the interference source, supplies an interference signal, which is analyzed in a frequency analysis circuit 21 with respect to the frequencies of the interference spectral lines. The frequency f1 of the first interference spectral line appears at the output of the frequency analysis circuit 21. This frequency is, for example, f1=15.45 Hz.
[0033] The Fourier transform circuit 13 has 256 frequency cells. The frequency resolution of the Fourier transform is determined by a frequency interval derived from the frequency f1. A frequency range of 256 Hz is sufficient for analyzing the received signal, so the frequency interval in this case would be 1 Hz. However, to match the frequency interval to the frequency f1 of the interference spectral line, the energy at f1=15.45 Hz should be written into frequency cell 15, not the energy component of the spectrum at 15 Hz. The frequency interval is selected to be f=1.03 Hz=15.45 / 15. The spectrum is then determined in frequency steps of 1.03 Hz. The coefficients of the Fourier analysis are stored in the Fourier transform circuit 13 in multiples of 1.03 Hz.The interference spectral line is then stored at 15·1.03=15.45 Hz in frequency cell 15, its first harmonic at 2·15·1.03=30.9 Hz in frequency cell 30, and its second harmonic at 3·15·1.03=46.35 Hz in frequency cell 45. The sampling frequency for the Fourier transform is selected from the number n = 256 frequency cells and the frequency interval to k·256·1.03 Hz, where k is an integer and greater than or preferably equal to two.
[0034] A divider 22 for determining the frequency interval Δf is connected downstream of the frequency analysis circuit 21. The divider 22 controls an arrangement 23 for forming a time window for the received signal, which is connected to the output side of the time gate 11. A frequency multiplier 24 is connected downstream of the divider 22, which provides the sampling frequency for the analog-to-digital converter 12 and the Fourier transformation circuit 13.
[0035] Fig. 3 shows the magnitude of the spectrum at the output of the Fourier transform circuit 13. The frequency axis is rasterized with the frequency interval 1.03 Hz, the interference spectral line at f1 can be seen in the frequency cell 15 at the frequency 15.45 Hz and also its harmonics at the frequencies 30.9 Hz and 46.35 Hz in the frequency cells 30 and 45. In a suppression circuit 30, which contains a smoothing circuit 32, the interference spectral lines at the output of the Fourier transform circuit 13 at the frequency f1 and its harmonics 2f1 and 3f1 are suppressed. The smoothing circuit 32 consists of a coefficient averager and an erase and overwrite arrangement which compares the coefficients at the frequencies f1, 2f1 and 3f in the frequency cells 15, 30 and 45 with coefficients at the frequency cells 14, ..., 16; 29, ..., 31; 44, ..., 46, forms the average value and writes the respective average value into the frequency cells 15, 30 and 45.
[0036] If only this single interference spectral line with its harmonics were present, the interference-free spectrum of the received signal would be available at the output 31 of the suppression circuit 30.
[0037] In the example considered here, the received signal still exhibits interference spectral lines at frequencies f2 and f3. In the next process step, the interference spectral line at frequency f2=25.5 Hz is eliminated. For this purpose, a control circuit 40 controls divider 22 and switches 51, 52, 53, and 54. A new frequency interval is formed from frequency f2 in divider 22. Switch 52 closes a feedback path at the output 31 of the suppression circuit 30 to the input of a computing circuit 33, to whose second input a new sampling frequency from the frequency multiplier 24 is supplied via switch 51.
[0038] The frequency f2 of the next interference spectral line to be eliminated is f2=25.5 Hz. In divider 22, the new frequency interval f=25.5:25=1.02 Hz is calculated, and in the frequency multiplier, the new sampling frequency f Abtast =k-256-1.02 Hz.
[0039] In the calculation circuit 33 in the masking circuit 30, the new frequency interval Δf=1.02 Hz and the previous frequency interval Δf=1.03 Hz and the new sampling frequency f Abtast =k-256-1.02 Hz, a sampled Si function is formed as an interpolation function, with which the first cleaned spectrum sent to the computing circuit 33 via the switch 52 is convolved. This process causes a new scaling of the frequency axis of the spectrum, as shown in Fig. 4. Coefficients of multiples of the new frequency interval Δf = 1.02 Hz are to be assigned to the frequency cells. Frequency cell 25 contains the interference spectral line with a frequency of 25.5. This interference spectral line is adjusted to the level curve of the spectrum in smoothing circuit 32.
[0040] The control circuit 40 controls the divider 22 and the suppression circuit again and the spectrum, which has been cleaned of the interference spectral lines at f1, 2f1, 3f1 and at f2, is switched from the output 31 of the suppression circuit 30 to the input of the arithmetic circuit 33. In the divider 22, the frequency f3 = 44 Hz of the next interference spectral line is used to calculate the new
[0041] Frequency interval Δf=44 / 44 Hz=1 Hz. In the frequency multiplier 24, a new, third sampling frequency f Abtast =k·256 Hz. In the computing circuit 33, the last cleaned spectrum is convolved with the new, sampled interpolation function. Fig. Figure 5 shows the spectrum, whose frequency scaling and resolution is 1 Hz and from which the interference spectral line at f3=44 Hz is smoothed in the smoothing circuit 32. The interference-free spectrum now only contains spectral components of the useful signal at 10 Hz, 20 Hz, 30 Hz and 45 Hz, as Fig. 5 shows.
[0042] Downstream of the masking circuit 30 is a computing circuit 60 for forming the inverse transform of the cleaned spectrum, which delivers the interference-free desired signal of the received signal on the output side. Any marginal interference that may occur at the beginning and end of the time signal is suppressed by a weighting circuit 61 with a window function, e.g., according to Hannig.
Claims
[1] Method for eliminating interference signals from received signals of a receiving system, wherein the received signals comprise useful signals and interference signals with distinct interference spectral lines, by detecting the frequencies of interference spectral lines and spectral masking, characterized by that a frequency interval which is a fraction of the frequency is derived from the frequency of at least one of the interference spectral lines by quotient formation, that a time window for a Fourier transformation of the received signal is selected to be equal to or a multiple of the reciprocal of the frequency interval and a sampling frequency is selected as an integer multiple of the frequency interval depending on the frequency range of the received signals, and that the spectrum of the received signal is leveled at the frequencies of the interference spectral lines and their harmonics. [2] Method according to claim 1, characterized bythat the frequency is divided by the nearest neighboring integer and the quotient is the frequency interval. [3] Method according to one of claims 1 to 2, characterized by that the mean value of those coefficients which are adjacent to frequencies of the spectral line and its harmonics is determined and the coefficients at the frequency of the interference spectral lines and their harmonics are replaced by the mean values. [4] Method according to one of claims 1 and 3 characterized by that the interference spectral lines are leveled one after the other. [5] Method according to claim 4, characterized bythat a first frequency interval is formed corresponding to the interference spectral line with the lowest frequency for selecting the time window and a first sampling frequency for creating a first spectrum of the received signals cleaned of this frequency and its harmonics, that the first cleaned spectrum is interpolated, that a further frequency interval and a sampling frequency dependent thereon is derived from the frequency of a further interference spectral line and the last cleaned, interpolated spectrum is sampled with the next sampling frequency and the associated interference spectral line and its harmonics are leveled and that interpolation, sampling and leveling are repeated until all interference spectral lines have been eliminated. [6] Method according to claim 5, characterized bythat interpolation and sampling are carried out by convolving the last cleaned spectrum with an interpolation function sampled at the next sampling frequency, the variable of which is the ratio of the newly determined and the last determined frequency interval. [7] Method according to claim 6, characterized by that an Si function is used for interpolation. [8] Method according to claim 7, characterized by that the Si function is multiplied by a shift factor. [9] Method according to one of claims 1 to 8, characterized by that the last cleaned spectrum is transformed back into the time domain and evaluated with a window function and provides the useful signal. [10] Device for carrying out the method according to one of claims 1 to 3 using a receiving system for detecting and / or direction finding with one or more transducers which supply the received signal individually and / or after direction formation, and a frequency analysis circuit for determining the frequencies of the interference spectral lines, characterized bythat a time gate (11) and, downstream of it, an analog / digital converter (12), a Fourier transformation circuit (13) and a suppression circuit (30) are provided for the received signal, into which a spectrum from the output of the Fourier transformation circuit (13) can be stored, that the suppression circuit (30) is designed to level or suppress coefficients at the frequencies of the interference spectral lines and their harmonics, that the frequency analysis circuit (21) is connected to control inputs of the suppression circuit (30), that a frequency divider (22) for forming a frequency interval (f) is arranged downstream of the frequency analysis circuit (21), the frequency divider (22) is connected to the time gate (11) and to a frequency multiplier (24) via an arrangement (23) for forming a time window, and that an output of the frequency multiplier (24) is connected to clock inputs of the analog / digital converter (12) and the Fourier transformation circuit (13). [11] Apparatus according to claim 10 for carrying out the method according to any one of claims 4 to 8, characterized by that the frequency divider (22) and the masking circuit (30) can be controlled by a control circuit (40), that the frequency divider (22) has frequency intervals (f) on the output side which are derived from the lowest and successively from higher frequencies of the interference spectral lines, that a computing circuit (33) for convolving a cleaned spectrum with an interpolation and sampling function is arranged in a feedback branch of the masking circuit (30) and can be controlled by the output of the frequency multiplier (24) and that the last cleaned spectrum can be switched successively to the computing circuit (33). [12] Device according to claim 11, characterized by that the masking circuit (30) contains a smoothing circuit with a coefficient averager and an erasing and overwriting arrangement for spectral lines of a spectrum. [13] Apparatus according to any one of claims 10 to 12 for carrying out the method according to claim 10, characterized by that the masking circuit (30) is followed by a computing circuit (60) for forming the inverse transform, to which an evaluation circuit (61) is connected.
Citation Information
Patent Citations
Method and device for determining the directions of incidence of wave energy
DE3017797A1
Control apparatus
US3599142A
Noise cancellation using tracking filter
US4232381A
Methods and device for attenuating spurious noise
US4594695A
Cited By
Instrument convolution removal spectrum overlapping peak decomposition method and device based on physical constraint
CN120507324A
