• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    A method of filtering a sampled digital input signal at a sampling frequency to obtain a filtered signal, the method comprising a first operation of applying a discrete Fourier transform to M points on a processed signal to obtain M spectrum points of the processed signal, each spectrum point of the processed signal corresponding to the even subscripts of a spectral analysis at 2 * M points of the processed signal and a second operation of applying a discrete Fourier transform at M points on the processed signal to obtain M spectrum points of the processed signal, each spectrum point of the processed signal corresponding to the odd indices of a spectral analysis at 2 * M points of the processed signal.
    公开号:FR3049131A1
    申请号:FR1600454
    申请日:2016-03-18
    公开日:2017-09-22
    发明作者:Jean Michel Hode
    申请人:Thales SA；
    IPC主号:
    专利说明:

    Method of filtering a digital input signal and associated filter
    The present invention relates to a method of filtering a digital input signal. The present invention also relates to a filter, a processing chain and an associated radar.
    For multiple applications in the field of radar, it is desirable to filter a digitized signal with a specific transfer function.
    For this purpose, transversal type filters with finite impulse response are used. Such filters are often referred to by the acronym FIR referring to the English terminology "Finite Impulse Response" which means "finite impulse response". FIR filters implement operations including the use of time delays of the signal, gains and summations. The number of operations is equal to the length of the impulse response of the FIR filter considered (the length being expressed in number of samples).
    However, when the length of the impulse response of the filter is very large, as is the case for the compression of pulses involved in the radars, the carrying out of the filtering becomes problematic or even impossible taking into account the very large number of operations involved.
    To circumvent such a problem, it is known to perform certain operations in the frequency space. For this, a Fourier transformation is applied to move from the time domain to the frequency domain, the filtering operation then becoming multiplicative, then a Fourier transformation is applied thereafter to return to the time domain.
    In practice, the time is divided into sequences and the Fourier transformation is implemented by a fast Fourier transform often designated by the acronym FFT for "Fast Fourier Transform" (Fast Fourier Transform in French). More precisely, the passage from the domain of the space of time to the space of the frequencies is obtained by use of an FFT whereas the passage from the domain of the space of the frequencies to the space of the times is obtained by the use of an IFFT. The acronym IFFT refers to "Inverse Fast Fourier Transform" (Fast Fourier Transform in French). The use of fast inverse Fourier transforms or not implies a size at least equal to the length of the filter. In fact, if the size of the fast Fourier transform inverse or not is strictly equal to the size K of the filter, then the process makes it possible to obtain only one point on K, the K-1 other calculated points n ' not being usable. If the size of the fast Fourier transform inverse or not is strictly equal to twice the size K of the filter, ie 2K, then the process makes it possible to obtain K points on 2K, the K other calculated points not being usable. Thus, by splitting the process, it is possible to calculate 2 K points on 2K and to access all the required points.
    However, this shows that half of the calculated points are lost, which increases the computing load and complicates the implementation of the filter.
    There is therefore a need for a filtering method of a digital input signal that is easier to implement.
    For this purpose, there is described a method of filtering a digital input signal sampled at a sampling frequency to obtain a filtered signal, the method comprising providing an input signal, signal transmission of input on two channels, obtaining a first output signal by implementing the first following operations on the first channel and obtaining a second output signal by implementing the following second operations on the second way. Obtaining a first output signal by carrying out the following first operations on the first channel: a first operation of first processing of the input signal to obtain a processed signal, a first operation of applying a first signal; discrete Fourier transform at M points on the processed signal to obtain M spectrum points of the processed signal, M being an integer strictly greater than 2, each spectrum point of the processed signal corresponding to the even indices of a spectral analysis at 2 * M points of the processed signal and being identified bijectively by an index k, where k is an even number between 0 and 2 * M-1, a first operation of second processing of the spectrum points of the processed signal to obtain first selected points , a first operation of applying the inverse discrete Fourier transform to M points on the first points selected to obtain a first sign al and a first third processing operation of the first signal to obtain a first output signal. Obtaining a second output signal by implementing the following second operations on the second channel: a second operation of first processing of the input signal to obtain a processed signal, a second operation of application of a second signal; discrete Fourier transform at M points on the processed signal to obtain M spectrum points of the processed signal, M being an integer strictly greater than 2, each spectrum point of the processed signal corresponding to the odd indices of a spectral analysis at 2 * M points of the signal processed and being identified bijectively by an index k, k being an odd number between 0 and 2 * M-1, a second operation of second processing of the spectrum points of the processed signal to obtain second selected points , a second operation of applying the inverse discrete Fourier transform to M points on the second points selected to obtain a d said second signal and a second third processing operation of the second signal to obtain a second output signal. The method also includes recombining the first output signal and the second output signal to obtain the filtered signal.
    According to particular embodiments, the method comprises one or more of the following characteristics, taken in isolation or in any technically possible combination: the second operation of first processing comprises the implementation of a frequency translation of a value equal to the ratio between the sampling frequency and the number 2 * M, and the second third processing operation comprises the implementation of a frequency translation applied to the second signal of a value equal to the opposite of the ratio between the sampling frequency and the number 2 * M. the first second processing operation comprises the implementation of the shift of the spectrum points of the processed signal of M samples to obtain offset points and the calculation of the sum of the spectrum points of the processed signal and the offset points, and the second second processing operation comprises implementing the shift of the spectrum points of the processed signal of M samples to obtain offset points and calculating the sum of the spectrum points of the processed signal and offset points. the first first processing operation comprises the implementation of the shift of the input signal of M samples to obtain an offset signal and the calculation of the sum of the input signal and the shifted signal, and the second operation of the first processing comprises implementing the shift of the input signal of M samples to obtain an offset signal and calculating the difference between the input signal and the shifted signal. the recombination step is implemented by calculating the difference between the first output signal and the second output signal. the recombination step is implemented by calculating the difference between the first output signal and the second output signal, to obtain a first calculation signal, calculating the sum of the first output signal and the second output signal; output, for obtaining a second intermediate signal of calculation, -shopping of the second intermediate signal for calculating M samples to obtain a second calculation signal, and calculating the sum of the first calculation signal and the second calculation signal to obtain the signal filtered.
    There is also described a filter adapted to filter a digital input signal sampled at a sampling frequency to obtain a filtered signal, the filter comprising an input terminal adapted to receive an input signal, a first channel specific to obtaining a first output signal by implementing first operations, a second path capable of obtaining a second output signal by the implementation of second operations, a transmitter capable of transmitting the input signal on the first channel and the second channel, a mixer capable of recombining the first output signal and the second output signal to obtain the filtered signal, the first channel being capable of implementing the following first operations: a first operation of the first signal processing; input to obtain a processed signal, a first operation of applying a discrete Fourier transform to M points on the processed signal to obtain M spectrum points of the processed signal, M being an integer strictly greater than 2, each spectrum point of the processed signal corresponding to the even subscripts of a spectral analysis at 2 * M points of the processed signal and being identified bijectively by an index k, k being an even number between 0 and 2 * M-1, a first operation of second processing of the spectrum points of the processed signal to obtain selected first points, a first operation of applying the inverse discrete Fourier transform to M points on the selected first points to obtain a first signal and a first third processing operation of the first signal to obtain a first output signal. The second channel is adapted to implement the following second operations: a second operation of first processing of the input signal to obtain a processed signal, a second operation of applying a discrete Fourier transform to M points on the signal processed to obtain M spectrum points of the processed signal, M being an integer strictly greater than 2, each spectrum point of the processed signal corresponding to the odd indices of a spectral analysis at 2 * M points of the processed signal and being spotted bijective by an index k, where k is an odd number between 0 and 2 * M-1, a second operation of second processing of the spectrum points of the processed signal to obtain selected second points, a second operation of applying the transform Discrete Fourier inverse to M points on the second selected points to obtain a second signal and a second operation third processing the second signal to obtain a second output signal. The description also relates to a processing chain comprising at least one filter as previously described.
    According to particular embodiments, the processing chain comprises one or more of the following characteristics, taken in isolation or in any technically possible combination: the processing chain is a programmable logic circuit. the processing chain is an integrated circuit specific to an application.
    In addition, there is also described a radar comprising a processing chain as previously described. Other features and advantages of the invention will appear on reading the following description of embodiments of the invention, given by way of example only and with reference to the drawings which are: FIG. 1, a diagrammatic view an example of a radar comprising several filters, - figure 2, a block diagram showing the operations performed by an example of a filter, - figure 3, a block diagram showing the operations performed by another example of a filter, - figure 4 , a block diagram showing the operations carried out by yet another example of a filter, and - FIG. 5, a block diagram illustrating the proper operation of the filter of FIG. 4.
    A radar 10 is schematically illustrated in FIG.
    The radar 10 is adapted to receive an input signal 10E and to convert the input signal 10E into a usable output signal 10S for subsequent uses.
    The radar 10 includes an antenna 12 and a processing chain 14. The antenna 12 is adapted to receive the input signal 10E.
    The processing chain 14 is suitable for converting the input signal 10E into a 10S output signal.
    The processing chain 14 is capable of filtering the input signal 10E.
    According to the example of Figure 1, the processing chain 14 comprises three filters 16, 18 and 20 in series.
    In fact, the first filter 16 comprises an input terminal 16E connected to the antenna 12 by a first wire 22 and an output terminal 16S connected to the second filter 18 by a second wire 24.
    The second filter 18 comprises an input terminal 18E connected to the output terminal 16S of the first filter 16 by the second wire 24 and an output terminal 18S connected to the third filter 20 by a third wire 26.
    The third filter 20 has an input 20E connected to the output terminal 18S of the second filter 18 by the third wire 26 and an output terminal 20S connected to the fourth wire 28 transmitting the output signal 10S.
    According to another embodiment, the processing chain 14 comprises a single filter.
    In a variant, the processing chain 14 comprises any number of filters, for example 2, 4 or 6.
    The processing chain 14 is, for example, a programmable logic circuit. Such a circuit is often referred to by the acronym FPGA, acronym for the expression "field-programmable grating array", a network of programmable gates in situ.
    In another example, the processing chain 14 is an integrated circuit specific to an application. Such a circuit is often referred to by the acronym ASIC (acronym for "application-specific integrated circuit", meaning "application-specific integrated circuit").
    Each filter 16, 18 and 20 is adapted to filter a digital input signal sampled at a sampling frequency to obtain a filtered signal.
    To simplify the description, it is assumed that each of the filters 16, 18 and 20 is identical.
    In a variant, each filter of the processing chain 14 is different.
    An example of second filter 18 is illustrated more precisely in FIG.
    The filter 18 comprises a first channel 181, a second channel 182, a transmitter 183 and a mixer 184.
    The input terminal 18E is adapted to receive an input signal.
    The first channel 181 is able to obtain a first output signal by implementing first operations.
    The first channel 181 is able to implement a first operation of the first processing of the input signal to obtain a processed signal.
    According to the example of Figure 2, the first processing is the sending of the input signal.
    The first channel 181 is also able to apply a discrete Fourier transform to M points on the processed signal to obtain M spectrum points of the processed signal, M being an integer strictly greater than 2, each spectrum point of the processed signal corresponding to even indices of a spectral analysis at 2 * M points of the processed signal and being identified bijectively by an index k, k being an even number between 0 and 2 * M-1. In this case, the points are numbered from 0 to 2 * M-1.
    This amounts to calculating the even coefficients of the spectral analysis at 2 * M points of the processed signal.
    For example, the computed discrete Fourier transform is a fast Fourier transform denoted FFTM.
    This is illustrated schematically in Figure 2 by a box in which it is written FFTMet by a multiplier on which arrives an arrow with the inscription "p2m".
    The first channel 181 is also suitable for implementing a first operation of second processing of the spectrum points of the processed signal to obtain selected first points.
    The first channel 181 is able to implement the shift of the spectrum points of the processed signal of M samples to obtain offset points and the calculation of the sum of the spectrum points of the processed signal and the offset points.
    The offset is illustrated schematically in FIG. 2 by a box in which it is written "z'M" with reference to an offset technique by using the Z-transform.
    A sum sign on which two arrows arrive, one corresponding to a path passing through the box in which it is written "z'M" and the other corresponding to a path that does not pass there schematically shows the first operation of second treatment.
    The first channel 181 is also suitable for implementing a first operation of applying the inverse discrete Fourier transform to M points on the first selected points to obtain a first signal.
    For example, the calculated discrete Fourier transform is a fast Fourier transform denoted IFFTM.
    This is illustrated schematically in FIG. 2 by a box in which it is written IFFTM.
    The first channel 181 is also suitable for implementing a first third processing operation of the first signal to obtain a first output signal.
    In this case, the first operation of the third treatment consists in transmitting the first output signal to the mixer 184.
    The second path 182 is able to obtain a second output signal by implementing second operations.
    The second channel 182 is capable of implementing a second operation of first processing of the input signal to obtain a processed signal.
    In this case, the second channel 182 is able to implement a frequency translation of a value equal to the ratio between the sampling frequency and the number 2 * M.
    Such an operation is symbolized by an arrow on a multiplier, arrow on which is written "ejTTn / M" by reference to a usual technique of translation which consists in multiplying the signal by a well-chosen complex exponential.
    The second channel 182 is adapted to implement a second operation of applying a discrete Fourier transform to M points on the processed signal to obtain M spectrum points of the processed signal, each spectrum point of the processed signal corresponding to the indices. odd spectral analysis at 2 * M points of the processed signal and being located bijectively by an index k, k being an odd number between 0 and 2 * M-1.
    The second channel 182 is suitable for implementing a second operation of second processing of the spectrum points of the processed signal to obtain selected second points.
    The second path 182 is thus able to implement the shift of the spectrum points of the processed signal of M samples to obtain offset points and the calculation of the sum of the spectrum points of the processed signal and offset points. From the point of view of the signal, the offset is a time delay.
    The offset is illustrated schematically in FIG. 2 by a box in which it is written "z'M" with reference to an offset technique by using the Z-transform.
    A sum sign on which two arrows arrive, one corresponding to a path passing through the box in which it is written "z'M" and the other corresponding to a path that does not pass there schematically shows the second operation of second treatment.
    The second path 182 is also adapted to implement a second operation of applying the inverse discrete Fourier transform to M points on the second selected points to obtain a second signal.
    The second path 182 is also suitable for implementing a second third processing operation of the second signal to obtain a second output signal.
    In this case, the second channel 182 is able to implement a frequency translation of a value opposite to the ratio between the sampling frequency and the number 2 * M.
    Such an operation is symbolized by an arrow on a multiplier, arrow on which is written "ejTTn / M" by reference to a common technique of translation which consists in multiplying the signal by a well-chosen complex exponential.
    The transmitter 183 is capable of transmitting the input signal on the first channel 181 and the second channel 182.
    The mixer 184 is able to recombine the first output signal and the second output signal to obtain the filtered signal.
    In this case, since the recombination is obtained by a difference of the first output signal and the second output signal, the mixer 184 is represented by a circle with a + sign and a - sign.
    Operation of the second filter 18 is now described with reference to an exemplary implementation of a method of filtering a digital input signal sampled at a sampling rate to obtain a filtered signal.
    The method comprises a supply step, a transmission step, a step of obtaining a first output signal, a step of obtaining a second output signal and a recombination step. In the supplying step, the input signal is supplied to the input terminal 18E of the second filter 18. At the transmission step, the input signal is transmitted by the transmitter 183 on the two channels 181 and 182. At the step of obtaining the first output signal, the fast M-point Fourier transform of the input signal is calculated to obtain the even-order coefficients of a spectral analysis at 2 * M points of the input signal. input signal.
    A first operation is then performed to implement the shift of the spectrum points of the processed signal of M samples to obtain offset points and to calculate the sum of the spectrum points of the processed signal and the offset points. First selected points are thus obtained.
    It is then applied an application of the inverse discrete Fourier transform to M points on the selected first points to obtain a first output signal.
    The first output signal is sent to the transmitter 183. At the step of obtaining the second output signal, it is implemented a frequency translation of a value equal to the ratio between the sampling frequency and the number 2 * M. A processed signal is thus obtained.
    The M-point fast Fourier transform of the processed signal is calculated to obtain the odd-order coefficients of a spectral analysis at 2 * M points of the input signal.
    A second operation is then performed to implement the shift of the spectrum points of the processed signal of M samples to obtain offset points and to calculate the sum of the spectrum points of the processed signal and the offset points. Second selected points are thus obtained.
    It is then applied an application of the inverse discrete Fourier transform to M points on the second selected points to obtain a second signal.
    It is then implemented an implementation of a frequency translation applied to the second signal of a value equal to the opposite of the ratio between the sampling frequency and the number 2 * M. This makes it possible to obtain an output signal. The recombination step is then implemented using the mixer 184 to obtain the filtered signal by making the difference between the first output signal and the second output signal.
    The method makes it possible to obtain a filtered signal more easily.
    In fact, the method, compared to a method of the state of the art, makes it possible to limit the required memory space by 40% as demonstrated by commenting on FIG. 5.
    In addition, it is also possible to pool the calculations since the two channels 181 and 182 are synchronous.
    In a variant, the second filter 18 is in accordance with the embodiment of FIG.
    The elements identical to the embodiment of FIG. 2 are not described again. Only the differences are highlighted.
    The second filter 18 according to the embodiment of FIG. 3 differs from the second filter 18 according to the embodiment of FIG. 2 by the operations that the first lanes 181 and the second lane 182 are suitable for carrying out and by the nature of the operations performed during the recombination step.
    In the case of FIG. 3, each of the first operations of first processing and of second processing consists of transmitting the signal considered.
    In the case of FIG. 3, the second operation of first processing comprises the implementation of a frequency translation of a value equal to the ratio between the sampling frequency and the number 2 * M.
    The second treatment consists in transmitting the signal considered.
    The second third processing operation comprises the implementation of a frequency translation applied to the second signal of a value equal to the opposite of the ratio between the sampling frequency and the number 2 * M. In the recombination step, the difference between the first output signal and the second output signal is calculated to obtain a first calculation signal.
    The sum of the first output signal and the second output signal is then calculated to obtain a second intermediate calculation signal. It is also implemented an offset of the second intermediate signal for calculating M samples to obtain a second calculation signal.
    It is also calculated the sum of the first calculation signal and the second calculation signal to obtain the filtered signal.
    The same advantages relate to the embodiment according to FIG.
    According to another variant, the second filter 18 is in accordance with the embodiment of FIG.
    The elements identical to the embodiment of FIG. 2 are not described again. Only the differences are highlighted.
    The second filter 18 according to the embodiment of Figure 4 differs from the second filter 18 according to the embodiment of Figure 2 by the operations that the first path 181 and second path 182 are adapted to implement.
    Similar elements
    In the case of FIG. 4, the first first processing operation implements the shift of the input signal of M samples to obtain an offset signal and the calculation of the sum of the input signal and the shifted signal.
    Each first operation of second treatment and third treatment consists of transmitting the signal.
    According to the example of FIG. 4, the second operation of the first processing implements the shift of the input signal of M samples to obtain an offset signal and the calculation of the sum of the input signal and the shifted signal. The second operation of first processing then implements a frequency translation of a value equal to the ratio between the sampling frequency and the number 2 * M.
    The second operation of the second processing consists of transmitting the signal.
    The second third processing operation implements a frequency translation applied to the second signal of a value equal to the opposite of the ratio between the sampling frequency and the number 2 * M.
    The same advantages relate to the embodiment according to FIG.
    In the following section, with reference to FIG. 5, it is shown that the proposed calculations make it possible to obtain the desired filtering.
    To calculate a slice of M output points, a slice of 2M points is used for indices n which, by convention, will be such that 0 <η <2M-1.
    To understand the rest of the calculation, it is necessary to explain generalities on the decomposition of a discrete Fourier transform (DFT) in Radix (where x is the number of factorizable points).
    The N-point discrete Fourier transform of a sequence of N points xn (N being an integer) is:
    In the case where N is the product N = Ni x N2 of two integers Ni and N2, it can be written:
    There are two ways to decompose the discrete Fourier transform, the two ways corresponding to the cascade of a discrete Fourier transform with Νί points and a discrete Fourier transform with N2 points. This is the order in which these discrete Fourier transforms are performed which differs. One of the discrete Fourier transforms involved is weighted by a local oscillator allowing fine shifting of the spectrum; the exponential term allowing this translation is called "tweedle factor":
    In principle, the output index varying the fastest is m ^ Since it corresponds to the discrete Fourier transform "fast" that is to say non-multiplexed. Thus, the order of output of the frequencies is not the natural order: indeed, leave first the m being worth 0 modulo Ni then the m being worth 1 and so on until the m being worth Ni - 1 always 0 modulo ISL; this order is called "bit reverse" because it corresponds to reversing the binary representation of the index to obtain the output rank (for a number of power points of 2).
    The realization of the inverse discrete Fourier transform, which recovers the frequency data in "reverse bit", is carried out in a dual manner, by reversal of the operations.
    According to a first case, the Ni-point discrete Fourier transform ("fast") is followed by the N2-point Fourier transform ("multiplexed"), which corresponds to:
    The realization of the Fourier transform is carried out as the cascade of a non-multiplexed Fourier transform and a N2 multiplexed Fourier transform. The tweedle factor is placed in front of the first one. Fourier transform stage.
    According to a second case, the N2-point discrete Fourier transform ("multiplexed") is followed by the Ni-Point Fourier transform ("fast"), which corresponds to:
    The realization is carried out this time as the cascade of a Ni-multiplexed Ni-point discrete Fourier transform and a non-multiplexed Ni-point discrete Fourier transform. The tweedle factor is placed between the two stages of discrete Fourier transform. The first discrete Fourier transform realized being multiplexed, this decomposition accommodates parallelized input data; the multiplexing is then transformed into parallel processing.
    Following one of the classical versions of decomposition in radix2, the FFT at 2M points of this sequence is written:
    where FFTm denotes the M-point FFT.
    It should be noted that this spectral density can also be written in two ways that will simplify the implementation:
    Noting pm the spectral response of the filter, it is obtained the response filtered by inverse transformation:
    The answer is calculated only for the points 0 <k <M-1 (due to the folding inherent to the FFT) to finally obtain:
    where IFFTm denotes the inverse FFT at M points.
    Formulas giving jb2m + p and Ck can be obtained from the previous block diagrams that allow convolutions with double size responses to be performed using the same size of FFT and IFFT, with storage on M points.
    The block diagram of FIG. 2 is then obtained. Such a block diagram corresponds to the first formulation of the spectral density b2m + p.
    By dragging the storage function to the output it can be reduced by half, which leads to the block diagram of Figure 3.
    Finally, the sliding of the memorization towards the input, which corresponds in fact to the second formulation of the spectral density b2m + p, leads to the block diagram of FIG. 4 which can be considered as a dual version of the block diagram of Figure 3.
    The transactions involved in these block diagrams correspond to the first rank (for the FFT) or the last rank (for the IFFT) of "tweedle factors" not taken into account, in principle, in the FFT at M points. Theoretically, these digital operations are to be reset at each FFT frame.
    However, it should be noted that this is actually not necessary in practice because, in the absence of reinitialization, the sign of these operations changes at each frame, so that the two sign changes compensate each other.
    It can be disassembled that such a filtering is more economical in terms of memory resources in the case where a realization in the stream is required.
    When the filtering corresponds to the filtering adapted to a given signal (a received pulse for example), the complex gain of the matched filter is the conjugate of the spectrum of the pulse. If this spectrum is obtained by FFT at 2M points of a pulse starting at n = 0 and ending before n = M, then the time response of the matched filter is then placed in the negative times.
    If rn is the pulse and if xn is the received signal, the matched filter performs the following operation, which is the calculation of the autocorrelation of the pulse when the received signal xn is equal to the pulse rn, the peak of correlation arriving at the beginning of the impulse:
    The conjugate of the complex gain of the matched filter is:
    Since the pulse is zero for the negative times, it is then obtained the result of two FFTs with M points:
    This gives the check block diagram of Figure 5 a more symmetrical view to directly verify that each implementation of the second filter 18 actually leads to the desired result.
    At the input of the IFFT of the first channel 181, it is obtained:
    It is obtained at the output of this IFFT of the first channel 181:
    For the second channel 182, at the input of the IFFT, it is obtained:
    These last equations are written:
    It is obtained at the output of this IFFT of the second channel 182, and after multiplication by the "tweedle factor":
    Now: • on the upper track:
    • on the lower track:
    The global output is:
    That is to say :
    Is :
    In fact, it has been shown that the desired result is obtained at a time shift of M points (ensuring the causality of the process) and at a multiplicative coefficient 2:
    This demonstration carried out for the block diagram of FIG. 4 is easily transposed for the other two block diagrams of FIGS. 2 and 3.
    In all cases, this reasoning shows that the major advantage of the proposed method is the saving of storage resources. In fact, an FFT requires a storage depth equal to its size.
    In the case of a method according to the state of the art, the memory space required is therefore equal to the resources required for an FFT (2 FFT and 2 IFFT), ie: 4 x 2M = 8M points. Conversely, in the case of the proposed method, the FFT being half size, the size required by the FFT and IFFT thus goes to 4M points to which must be added M points to realize the specific storage carried out at the head; this corresponds to a calculation of 5M points. The savings achieved is close to 40%.
    The second advantage of the proposed method lies in the fact that the two channels 181 and 182 are synchronous while these channels are shifted by one half-frame in the case of a method according to the state of the art. In fact, if the tweedle factors of the FFT and the IFFT are calculated instead of being stored in tables (which is desirable for large size FFTs), this makes it possible to pool this calculation between the two channels 181 and 182 of the process, whereas this was not possible in a method according to the state of the art.
    In other words, it has been shown that the proposed method allows easier implementation of a filtering of a digital input signal.
    权利要求:
    Claims (10)
    [1" id="c-fr-0001]
    A method of filtering a sampled digital input signal at a sampling frequency to obtain a filtered signal, the method comprising: - supplying an input signal, - transmitting the input signal. on two channels (181, 182), - obtaining a first output signal by implementing the following first operations on the first channel (181): • first processing of the input signal to obtain a processed signal, • applying a discrete Fourier transform to M points on the processed signal to obtain M spectrum points of the processed signal, M being an integer strictly greater than 2, each spectrum point of the processed signal corresponding to the even subscripts of an analysis spectral at 2 * M points of the processed signal and being identified bijectively by an index k, k being an even number between 0 and 2 * M-1, • second processing of the spectrum points of the signal signal to obtain first selected points, • application of the inverse discrete Fourier transform to M points on the first points selected to obtain a first signal, • third processing of the first signal to obtain a first output signal, - obtaining of a second output signal by the implementation of the following second operations on the second channel (182): • first processing of the input signal to obtain a processed signal, • application of a discrete Fourier transform to M points on the processed signal to obtain M spectral points of the processed signal, M being an integer strictly greater than 2, each spectrum point of the processed signal corresponding to the odd indices of a spectral analysis at 2 * M points of the processed signal and being spotted from bijective way by an index k, k being an odd number between 0 and 2 * M-1, • second treatment of the spectrum points of the signal t to obtain selected second points, • application of the discrete inverse Fourier transform to M points on the second points selected to obtain a second signal, • third processing of the second signal to obtain a second output signal, - the recombination of the first signal. output signal and the second output signal to obtain the filtered signal.
    [2" id="c-fr-0002]
    2. - A filtering method according to claim 1, wherein: the second operation of first processing comprises the implementation of a frequency translation of a value equal to the ratio between the sampling frequency and the number 2 * M and the second third processing operation comprises the implementation of a frequency translation applied to the second signal of a value equal to the opposite of the ratio between the sampling frequency and the number 2 * M.
    [3" id="c-fr-0003]
    3. - A filtering method according to claim 1 or 2, wherein: the first second processing operation comprises the implementation of the offset of the spectrum points of the processed signal of M samples to obtain offset points and the calculation of the sum of the spectrum points of the processed signal and the offset points, and the second second processing operation comprises the implementation of the shift of the spectrum points of the processed signal of M samples to obtain offset points and the calculation of the sum of the spectrum points of the processed signal and offset points.
    [4" id="c-fr-0004]
    4. A filtering method according to claim 1 or 2, wherein: the first operation of first processing comprises implementing the shift of the input signal of M samples to obtain an offset signal and the calculation of the sum of the input signal and the shifted signal, and the second first processing operation comprises implementing the shift of the input signal of M samples to obtain an offset signal and calculating the difference between the input signal and the signal shifted.
    [5" id="c-fr-0005]
    5. - The filtering method according to any one of claims 1 to 4, wherein the recombination step is implemented by calculating the difference between the first output signal and the second output signal.
    [6" id="c-fr-0006]
    6. - A filtering method according to claim 1 or 2, wherein the recombination step is implemented by: - calculating the difference between the first output signal and the second output signal, to obtain a first signal of calculation, - calculation of the sum of the first output signal and the second output signal, to obtain a second intermediate calculation signal, - offset of the second intermediate calculation signal of M samples to obtain a second calculation signal, and - calculation the sum of the first calculation signal and the second calculation signal to obtain the filtered signal.
    [7" id="c-fr-0007]
    7, - filter (16, 18, 20) adapted to filter a digital input signal sampled at a sampling frequency to obtain a filtered signal, the filter (16, 18, 20) comprising: - an input terminal (16E, 18E and 20E) adapted to receive an input signal, - a first channel (181) capable of obtaining a first output signal by the implementation of first operations, - a second channel (182) suitable for obtaining a second output signal by the implementation of second operations, - a transmitter (183) capable of transmitting the input signal on the first channel (181) and the second channel (182), - a mixer (184) clean recombining the first output signal and the second output signal to obtain the filtered signal, the first channel (181) being able to carry out the following first operations: • first processing of the input signal to obtain a processed signal, • application of a transformation of e Fourier discrete at M points on the processed signal to obtain M spectrum points of the processed signal, M being an integer strictly greater than 2, each spectrum point of the processed signal corresponding to the even indices of a spectral analysis at 2 * M points of the processed signal and being identified bijectively by an index k, k being an even number between 0 and 2 * M-1, • second processing of the spectrum points of the processed signal to obtain selected first points, • application of the inverse discrete Fourier transform at M points on the selected first points to obtain a first signal, • third processing of the first signal to obtain a first output signal, and the second channel (182) being suitable for implementing the second operations following: • first processing of the input signal to obtain a processed signal, • application of a discrete Fourier transform to M points on the e processed signal to obtain M spectrum points of the processed signal, M being an integer strictly greater than 2, each spectrum point of the processed signal corresponding to the odd indices of a spectral analysis at 2 * M points of the processed signal and being spotted bijectively by an index k, where k is an odd number between 0 and 2 * M-1, • second processing of the spectrum points of the processed signal to obtain selected second points, • application of the inverse discrete Fourier transform to M points on the second points selected to obtain a second signal, • third processing of the second signal to obtain a second output signal.
    [8" id="c-fr-0008]
    8. - treatment chain (14) comprising at least one filter (16, 18, 20) according to claim 7.
    [9" id="c-fr-0009]
    9. - processing chain according to claim 8, the processing chain (14) being a programmable logic circuit or an integrated circuit specific to an application.
    [10" id="c-fr-0010]
    10. - Radar (10) comprising a treatment chain (14) according to claim 8 or 9.
    类似技术:
    公开号 | 公开日 | 专利标题
    FR3049131A1|2017-09-22|METHOD FOR FILTERING A DIGITAL INPUT SIGNAL AND FILTER THEREFOR
    EP0190796B1|1989-12-20|System for signal analysis and synthesis filter banks
    EP2484075B1|2015-06-03|Systems for the multicarrier transmission of digital data and transmission methods using such systems
    EP2681889B1|2016-01-06|Method for modulating an oqam multi-carrier signal, and corresponding computer program and modulator
    FR2935856A1|2010-03-12|DIGITAL MULTI-CHANNEL SIGNAL TRANSMISSION SYSTEM USING FILTER BENCHES AND MEMORY PRELOADING FOR INITIALIZATION
    EP0081437A1|1983-06-15|Echo cancelling system
    FR2584213A1|1987-01-02|DEVICE FOR CALCULATING A DISCRETE, SLIDING FOURIER TRANSFORMATION, AND ITS APPLICATION TO A RADAR SYSTEM.
    EP1216557B1|2009-06-24|Method for transmitting an offset modulated biorthogonal multicarrier signal |
    EP0312463B1|1995-02-01|Digital filter device and radar comprising such a device
    FR2969804A1|2012-06-29|IMPROVED FILTERING IN THE TRANSFORMED DOMAIN.
    EP1940023A2|2008-07-02|Bank of cascadable digital filters, and reception circuit including such a bank of cascaded filters
    EP0688476B1|1997-05-21|Digital time signal filtering method and device for transmission channel echo correction
    EP3809287A1|2021-04-21|Method for simplifying a filter and associated devices
    EP3809286A1|2021-04-21|Method for filtering with reduced latency and associated devices
    EP0285495B1|1993-08-25|Digital signal processing system using a filter bank
    EP3809288A1|2021-04-21|Method for filtering with zero latency and associated devices
    EP1101283B1|2002-07-03|Method for producing digital nyquist filters with null interference between symbols and corresponding filtering device
    FR2817100A1|2002-05-24|METHOD FOR OPTIMIZING THE PERFORMANCE OF A TRANSMITTER FOR A MOBILE RADIO COMMUNICATION SYSTEM
    KR20120129003A|2012-11-28|Method and system for performing complex sampling of signals by using two or more sampling channels and for calculating time delays between these channels
    EP0763892B1|2001-05-09|Perfect reconstruction lattice filter bank
    WO2016097631A1|2016-06-23|Transmission method and device, and reception method and device
    WO2014166538A1|2014-10-16|Method and device for compensation of chromatic dispersion
    FR2915645A1|2008-10-31|Digital signal e.g. quadrature amplitude modulation signal, processing e.g. modulating, component for e.g. potable computer, has control unit configuring connections of ports to assure mathematic transformation function of signal
    同族专利:
    公开号 | 公开日
    US10598765B2|2020-03-24|
    US20170269192A1|2017-09-21|
    GB201704193D0|2017-05-03|
    FR3049131B1|2018-04-06|
    DE102017105808A1|2017-09-21|
    GB2550649A|2017-11-29|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US4992967A|1987-10-16|1991-02-12|Thomson-Csf|Digital filter using fourier transformation|
    US6324559B1|1998-10-16|2001-11-27|Telefonaktiebolaget Lm Ericsson |Odd-transform fast convolution|
    US20100235419A1|2009-03-13|2010-09-16|Yuki Yamamoto|Filtering apparatus, filtering method, program, and surround processor|EP3809286A1|2019-10-16|2021-04-21|Thales|Method for filtering with reduced latency and associated devices|
    EP3809287A1|2019-10-16|2021-04-21|Thales|Method for simplifying a filter and associated devices|
    EP3809288A1|2019-10-16|2021-04-21|Thales|Method for filtering with zero latency and associated devices|US4412325A|1981-07-13|1983-10-25|Telettra- Telefonia Elettronica E Radio S.P.A.|Equipment for single band multiplexing through digital processing|
    DE3610195C2|1986-03-26|1988-12-15|Ant Nachrichtentechnik Gmbh, 7150 Backnang, De|
    US5270953A|1991-05-23|1993-12-14|Rockwell International Corporation|Fast convolution multiplier|
    US7430258B2|2004-12-29|2008-09-30|Syracuse Research Corporation|Architecture for multi-channel digital signal processing|
    WO2007100666A2|2006-02-22|2007-09-07|University Of Akron|Interleaved method for parallel implementation of the fast fourier transform|
    EP1879293B1|2006-07-10|2019-02-20|Harman Becker Automotive Systems GmbH|Partitioned fast convolution in the time and frequency domain|
    FR2915038B1|2007-04-12|2012-08-03|Univ Paris Curie|HIGH FREQUENCY RECEIVER WITH MULTI-CHANNEL DIGITAL PROCESSING|
    EP2144172A1|2008-07-07|2010-01-13|Mitsubishi Electric R&D Centre Europe B.V.|Computation module to compute a multi radix butterfly to be used in DTF computation|
    JP2010127645A|2008-11-25|2010-06-10|Fuji Heavy Ind Ltd|Arbitrary signal generator|
    US8516027B2|2010-04-30|2013-08-20|Src, Inc.|Method and system for bit stacked fast Fourier transform|
    US8548100B2|2011-05-09|2013-10-01|Elta Systems Ltd.|Method and system for performing complex sampling of signals by using two or more sampling channels and for calculating time delays between these channels|CN109951174B|2019-04-16|2021-07-13|上海交通大学|Fast Fourier transform filtering method|
    法律状态:
    2017-03-31| PLFP| Fee payment|Year of fee payment: 2 |
    2017-09-22| PLSC| Publication of the preliminary search report|Effective date: 20170922 |
    2018-03-30| PLFP| Fee payment|Year of fee payment: 3 |
    2020-03-31| PLFP| Fee payment|Year of fee payment: 5 |
    2021-03-30| PLFP| Fee payment|Year of fee payment: 6 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1600454A|FR3049131B1|2016-03-18|2016-03-18|METHOD FOR FILTERING A DIGITAL INPUT SIGNAL AND FILTER THEREFOR|
    FR1600454|2016-03-18|FR1600454A| FR3049131B1|2016-03-18|2016-03-18|METHOD FOR FILTERING A DIGITAL INPUT SIGNAL AND FILTER THEREFOR|
    US15/458,222| US10598765B2|2016-03-18|2017-03-14|Method for filtering a numerical input signal and associated filter|
    GB1704193.0A| GB2550649A|2016-03-18|2017-03-16|Method for filtering a numberical input signal and associated filter|
    DE102017105808.7A| DE102017105808A1|2016-03-18|2017-03-17|Filtering method of a digital input signal and associated filter|
    [返回顶部]