Radar device and signal processing program

Abstract

To reduce influence of an interference wave in a reception signal better than ever before under an environment where radio wave interference occurs in a radar device, and to suppress a processing load for reduction.SOLUTION: Parameters of a mathematical formula representing a waveform of an interference wave are estimated as interference beginning time Tb, chirp slope K, amplitude A and a phase θ in an interference wave. The parameter estimation includes OMP processing that calculates a coefficient vector x in an arithmetic equation, in which an N rows and one column receiving waveform vector y corresponding to a wave form of a receiving wave is indicated by a product between an N rows and M columns dictionary matrix D being a set of basis vectors corresponding to a combination between the interference beginning time Tb and the chirp slope K, and an M rows and one column coefficient vector x including the amplitude A and the phase θ, to include only one non-zero component, using an OMP. And, before the OMP processing, an approximate calculation value of the chirp slope K is calculated, and a range of a value of the chirp slope K used in the basis vector of the dictionary matrix D is limited to be included in a predetermined range including the calculated approximate calculation value.SELECTED DRAWING: Figure 21

Description

[Technical field]

[0001] The present invention relates to a radar device and a signal processing program executed in the radar device. [Background technology]

[0002] Conventionally, radar systems are known that are mounted on vehicles and detect targets such as other vehicles and stationary objects around the vehicle. Such radar systems may cause radio interference with radar systems mounted on other vehicles. In the case of FMCW radars and the like that extract various information using the Fourier analysis results of beat signals, radio interference causes the noise floor of the frequency spectrum obtained by Fourier transforming the time waveform to rise. As a result, peaks based on the waves reflected from the target are buried in the noise floor, making it difficult to detect the target. FMCW stands for Frequency Modulated Continuous Wave.

[0003] In order to realize stable operation of a radar device even when radio interference as described above occurs, various proposals have been made in the past. For example, a radar device described in Patent Document 1 includes an object information calculation unit that calculates object information based on a received signal, an interference status analysis unit that measures the surrounding radio interference status between a predetermined minimum frequency and a maximum frequency during a search and modulation operation period that does not overlap with a modulation operation period, and a band selection unit that selects a frequency band of a transmission signal based on the radio interference status measured by the interference status analysis unit. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 6744481 Summary of the Invention [Problem to be solved by the invention]

[0005] As described above, the radar device described in Patent Document 1 measures the radio wave interference situation around the vehicle, searches for a frequency band with less interference, and selects the frequency band as the frequency band for the subsequent transmission signal. This aims to achieve stable operation of the radar device even when radio wave interference occurs. However, for example, when there are many other vehicles around the vehicle, it may be difficult to search for a frequency band with less interference.

[0006] On the other hand, it is also possible to realize stable operation of the radar device by reducing the influence of interference waves on the received signal through signal processing of the received signal. Various proposals have been made for this type of technology. The present invention has been made in consideration of the above-mentioned circumstances. That is, the present invention aims to provide a technology that can reduce the influence of interference waves on the received signal more effectively than ever before in an environment where radio wave interference occurs, and to suppress the processing load for the reduction.

[0007] In view of the above, an object of the present invention is to provide a method for preventing the above-mentioned problems. [Means for solving the problem]

[0008] In order to achieve the above object, the present invention provides a radar device (1), a transmission signal generating unit (51) for generating a transmission signal (TX) which is an electrical signal having a predetermined frequency change corresponding to a frequency modulation mode of a transmission wave, the transmission wave being a frequency-modulated radar wave, so as to transmit the transmission wave from a transmission antenna (2); The radar wave received by the receiving antenna (3) is mixed with a local signal (LO) to produce a predetermined cutoff frequency F LPS a signal processing unit (7) that performs an interference suppression process (S1800 to S1809) for suppressing the influence of an interference wave included in the received wave based on a signal processing result of the beat signal obtained by filtering the received wave; Equipped with The interference suppression process includes: A waveform estimation process (S1800 to S1808) for estimating a waveform of the interference wave; A subtraction process (S1809) of subtracting the estimated waveform of the interference wave from the waveform of the received wave; The waveform estimation process includes: a parameter estimation process (S1800 to S1807) for estimating parameters in an equation expressing a waveform of the interference wave using as parameters an interference start time Tb at which the interference wave starts to ride on the beat signal, a chirp slope K which is a slope of a time change in the frequency of the interference wave, an amplitude A of the interference wave, and a phase θ of the interference wave; an interference waveform calculation process (S1808) for calculating a waveform of the interference wave based on the parameters estimated by the parameter estimation process; Including, The parameter estimation process includes: an OMP process (S1802 to S1806) for selecting one of the basis vectors that maximizes an inner product with the received waveform vector y from the dictionary matrix D, in order to calculate the coefficient vector x in an arithmetic equation y=Dx, which is represented by the product of a dictionary matrix D of N rows and M columns that is a set of basis vectors corresponding to a combination of the interference start time Tb and the chirp slope K, and a coefficient vector x of M rows and 1 column that includes the amplitude A and the phase θ, using an orthogonal matching pursuit algorithm so that the coefficient vector x includes only one non-zero component and the rest are zero components; The radar device includes pre-processing (S1800, S1801, S110 to S170) for calculating an approximate value of the chirp slope K before the OMP processing, and limiting the range of values ​​of the chirp slope K used in the basis vectors of the dictionary matrix D so that the range falls within a predetermined range including the calculated approximate value.

[0009] In addition, the invention described in claim 11 is a method for transmitting a radar wave having a frequency modulated by a transmitting antenna (2), in which a local signal (LO) is mixed with a radar wave received by a receiving antenna (3), and a predetermined cutoff frequency F LPSA program used in a signal processing unit (7) of a radar device that obtains a beat signal by filtering the signal causing the signal processing unit to function to execute an interference suppression process (S1800 to S1809) for suppressing an influence of an interference wave included in the received wave based on a signal processing result for the beat signal; The interference suppression process includes: A waveform estimation process (S1800 to S1808) for estimating a waveform of the interference wave; A subtraction process (S1809) of subtracting the estimated waveform of the interference wave from the waveform of the received wave; The waveform estimation process includes: a parameter estimation process (S1800 to S1807) for estimating parameters in an equation expressing a waveform of the interference wave using as parameters an interference start time Tb at which the interference wave starts to ride on the beat signal, a chirp slope K which is a slope of a time change in the frequency of the interference wave, an amplitude A of the interference wave, and a phase θ of the interference wave; an interference waveform calculation process (S1808) for calculating a waveform of the interference wave based on the parameters estimated by the parameter estimation process; Including, The parameter estimation process includes: an OMP process (S1802 to S1806) for selecting one of the basis vectors that maximizes an inner product with the received waveform vector y from the dictionary matrix D, in order to calculate the coefficient vector x in an arithmetic equation y=Dx, which is represented by the product of a dictionary matrix D of N rows and M columns that is a set of basis vectors corresponding to a combination of the interference start time Tb and the chirp slope K, and a coefficient vector x of M rows and 1 column that includes the amplitude A and the phase θ, using an orthogonal matching pursuit algorithm so that the coefficient vector x includes only one non-zero component and the rest are zero components; and pre-processing (S1800, S1801, S110 to S170) for calculating an approximate value of the chirp slope K before the OMP processing, and limiting the range of values ​​of the chirp slope K to be used in the basis vectors of the dictionary matrix D so that it falls within a predetermined range including the calculated approximate value.

[0010] In this way, the radar device uses an orthogonal matching pursuit algorithm to calculate the coefficient vector x in the calculation formula y=Dx so that it contains only one non-zero component and the rest are zero components. This reduces the possibility of removing not only the interference wave but also the desired wave.

[0011] Furthermore, the radar device calculates an approximate value of the chirp slope K before the OMP process, and based on the approximate value, limits the range of values ​​of the chirp slope K used in the basis vectors of the dictionary matrix D. In this way, it becomes possible to effectively remove interference waves with fewer OMP processes, compared to a case in which basis vectors are prepared for the entire theoretically possible range of the chirp slope K.

[0012] The reference symbols in parentheses attached to each component indicate an example of the correspondence between the component and specific components described in the embodiments described below. [Brief description of the drawings]

[0013] [Figure 1] 1 is a block diagram showing a schematic configuration of a radar device according to an embodiment; [Diagram 2] 11 is a distance map showing a comparison of the Fourier transform results of the received waveform when there is interference and when there is no interference. [Diagram 3] 2 is a diagram showing an example of output signals in the in-phase signal AD converter and the quadrature phase signal AD converter shown in FIG. 1 in the form of time waveforms. [Figure 4] 1A and 1B are diagrams illustrating changes over time in frequency of a local signal, a received signal (before frequency downmixing), and a received signal (after frequency downmixing). [Diagram 5] 5 is a diagram showing a time change in beat frequency of an interference wave having the frequency characteristic shown in FIG. 4. [Figure 6] 1 is a graph showing changes over time in the real part, imaginary part, and absolute value of a beat signal. [Figure 7] FIG. 11 is a conceptual diagram showing an overview of a waveform estimation process when an orthogonal matching pursuit algorithm is used as a generally known method as a comparative example. [Figure 8] FIG. 13 is a conceptual diagram showing an example of a waveform estimation process using an orthogonal matching pursuit algorithm in a further comparative example. [Figure 9] FIG. 13 is a conceptual diagram showing an example of a waveform estimation process using an orthogonal matching pursuit algorithm in a further comparative example. [Figure 10] FIG. 13 is a conceptual diagram showing an example of a waveform estimation process using an orthogonal matching pursuit algorithm in a further comparative example. [Figure 11] FIG. 13 is a conceptual diagram showing an example of a waveform estimation process using an orthogonal matching pursuit algorithm in a further comparative example. [Figure 12] FIG. 13 is a conceptual diagram showing an example of a waveform estimation process using an orthogonal matching pursuit algorithm in a further comparative example. [Figure 13] FIG. 13 is a conceptual diagram showing an example of a waveform estimation process using an orthogonal matching pursuit algorithm in a further comparative example. [Figure 14] 13 is a table illustrating how the value of chirp slope K converges through waveform estimation processing using the orthogonal matching pursuit algorithm in an additional comparative example. [Figure 15] 13 is a table illustrating a state in which the value of the interference start time Tb converges by a waveform estimation process using the orthogonal matching pursuit algorithm in a further comparative example. [Figure 16A] 13A and 13B are diagrams illustrating how interference waves are removed from received waves according to an additional comparative example. [Figure 16B] 13A and 13B are diagrams illustrating how interference waves are removed from received waves according to an additional comparative example. [Figure 17] 11 is a table illustrating a state in which the value of chirp slope K converges by the waveform estimation process using the orthogonal matching pursuit algorithm in the first embodiment. [Figure 18] 11 is a table illustrating a state in which the value of interference start time Tb converges by the waveform estimation process using the orthogonal matching pursuit algorithm in the first embodiment. [Figure 19] FIG. 4 is a diagram showing how an interference wave is removed from a received wave according to the first embodiment. [Figure 20] FIG. 11 is a diagram showing a reduction in the noise floor in a distance FFT spectrum according to the first embodiment. [Figure 21] 4 is a flowchart of a process executed by a signal processing unit in the first embodiment. [Figure 22] 10 is a flowchart of a process executed by a signal processing unit in the second embodiment. [Diagram 23] This is an example of a two-dimensional map obtained by short-time Fourier transform. [Figure 24] 13 is a flowchart of a first type pre-processing. [Diagram 25] 13 is a flowchart of a second type pre-processing. [Figure 26] 13 is a flowchart of a third type pre-processing. [Figure 27] 13 is a diagram showing how an interference wave is removed from a received wave according to the second embodiment. FIG. [Figure 28] 13 is a flowchart of a process executed by a signal processing unit in the third embodiment. [Figure 29] An example of a two-dimensional map of two interference waves obtained by short-time Fourier transform. [Diagram 30] An example of a 2D map with one interference wave obtained by short-time Fourier transform. [Diagram 31] This is an example of a 2D map obtained by short-time Fourier transform after removing interference waves. [Diagram 32] 13A and 13B are diagrams illustrating how an interference wave is removed from a received wave according to the third embodiment. [Figure 33A] 11A and 11B are diagrams illustrating how interference waves are removed from received waves in a comparative example. [Figure 33B] 11A and 11B are diagrams illustrating how interference waves are removed from received waves in a comparative example. [Diagram 34] FIG. 13 is a diagram illustrating output timing of a local signal in the fourth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that various modified examples applicable to one embodiment may be hindered from understanding the embodiment if they are inserted in the middle of a series of explanations related to the embodiment. For this reason, the modified examples are described together after the explanation of the embodiment.

[0015] (First embodiment)

[0016] (composition) First, the first embodiment will be described. The radar device 1 shown in FIG. 1 is configured to be mounted on a vehicle and detect targets such as other vehicles and stationary objects present around the vehicle. The vehicle mounted with the radar device 1 according to this embodiment shown in FIG. 1 will be referred to as the "own vehicle" hereinafter. The radar device 1 mounted on the own vehicle will be referred to as the "own device" hereinafter. The radar device 1 according to this embodiment includes a transmitting antenna 2, a receiving antenna 3, and a control unit 4. The transmitting antenna 2 is provided to transmit a transmitting wave, which is a radar wave in a microwave band or a millimeter wave band, toward an external space of the own vehicle. The receiving antenna 3 is provided to receive a receiving wave, which is a radar wave in a frequency band corresponding to the transmitting wave. That is, the radar device 1 transmits a transmitting wave included in a predetermined frequency band toward an external space and receives a receiving wave included in the frequency band from the external space. The control unit 4 is electrically connected to the transmitting antenna 2 and the receiving antenna 3. In this embodiment, the radar device 1 has a configuration as a so-called FMCW radar. That is, the radar device 1 is configured to transmit and receive frequency-modulated radar waves. Each of the transmitting antenna 2 and the receiving antenna 3 may be an array antenna in which a plurality of antenna elements are arranged at different positions.

[0017] The control unit 4 includes a transmission circuit section 5, a reception circuit section 6, and a signal processing section 7. The transmission circuit section 5 is electrically connected to the transmission antenna 2 and configured to transmit a transmission wave from the transmission antenna 2. The reception circuit section 6 is electrically connected to the reception antenna 3 and configured to perform various signal processing such as amplification and detection processing on an RF signal generated by receiving a reception wave at the reception antenna 3. The signal processing section 7 is electrically connected to the transmission circuit section 5 and the reception circuit section 6 so as to be able to communicate signals. The signal processing section 7 is provided to control the operation of the transmission circuit section 5 to control the frequency and transmission timing of the transmission wave, and to execute various processes including an interference suppression process to be described later based on the signal processing result by the reception circuit section 6.

[0018] The transmission circuit unit 5 includes a transmission signal generating unit 51, a distributor 52, and a transmission side amplifier 53. The transmission signal generating unit 51 is provided to generate a transmission signal TX for transmitting a transmission wave, which is a frequency-modulated radar wave, from the transmission antenna 2, and output the transmission signal TX to the distributor 52. The transmission signal TX is an electric signal having a predetermined frequency change corresponding to the frequency modulation mode of the transmission wave. Specifically, the transmission signal generating unit 51 includes a phase synchronization control circuit (i.e., a PLL circuit), a voltage controlled oscillator (i.e., a VCO), and the like. PLL stands for Phase Locked Loop. VCO stands for Voltage Control Oscillator. Note that a configuration for generating a transmission signal TX using a phase synchronization control circuit or a voltage controlled oscillator is already publicly known or has been well known at the time of filing of the present application. For example, see JP2010-71899A, JP2011-12960A, JP2016-219938A, and the like. Therefore, further detailed description of the configuration of the transmission signal generating unit 51 is omitted in this specification. The distributor 52 is interposed in the signal transmission path between the transmission signal generating unit 51 and the transmission side amplifier 53, and is configured to distribute the power of the transmission signal TX output from the transmission signal generating unit 51 to the transmission side amplifier 53 and the receiving circuit unit 6. A signal having the same frequency characteristics as the transmission signal TX and distributed to the receiving circuit unit 6 by the distributor 52 is hereinafter referred to as a "local signal LO". The configuration of distributing the local signal LO from the transmission signal TX by the distributor 52 is also already publicly known or well known at the time of filing of this application, so further detailed description is omitted in this specification. The transmission side amplifier 53 is interposed in the signal transmission path between the distributor 52 and the transmitting antenna 2, and is configured to power amplify the transmission signal TX and output it to the transmitting antenna 2.

[0019] The receiving circuit unit 6 includes a receiving amplifier 61, a phase shifter 62, an in-phase signal mixer 63, a quadrature signal mixer 64, an in-phase signal filter 65, a quadrature signal filter 66, an in-phase signal AD converter 67, and a quadrature signal AD converter 68. The receiving amplifier 61 is electrically connected to the receiving antenna 3. The receiving amplifier 61 is provided to generate a receiving signal RX by amplifying an RF signal generated by the receiving antenna 3 receiving a receiving wave. In other words, the receiving signal RX is an electrical signal corresponding to the receiving wave, which is a radar wave received by the receiving antenna 3.

[0020] The phase shifter 62 is interposed in a signal transmission path between the distributor 52 and the quadrature signal mixer 64, and is configured to shift the phase of the local signal LO by −90 degrees. The in-phase signal mixer 63 and the quadrature signal mixer 64 are arranged in parallel at the output destination of the reception signal RX from the reception side amplifier 61. The in-phase signal mixer 63 and the quadrature signal mixer 64 are configured to perform so-called quadrature detection processing on the reception signal RX based on the transmission signal TX, i.e., the local signal LO. Specifically, the in-phase signal mixer 63 is arranged to mix the local signal LO and the reception signal RX to generate an I component, which is the real part of the IF signal corresponding to the frequency difference between the two, and output the I component to the in-phase signal filter 65. The IF signal is an abbreviation for an intermediate frequency signal. The quadrature signal mixer 64 is arranged to mix the local signal LO, whose phase has been shifted by the phase shifter 62, with the reception signal RX to generate a Q component, which is the imaginary part of the IF signal, and output the Q component to the quadrature signal filter 66. In addition, the above-mentioned configuration for generating the I and Q components of an IF signal using the local signal LO and a signal obtained by shifting the phase of the local signal LO has already been publicly known or well-known at the time of filing of the present application. For example, see JP 2021-47111 A, JP 2014-153216 A, JP 2001-91639 A, etc. Therefore, further detailed description of the phase shifter 62, the in-phase signal mixer 63, and the quadrature phase signal mixer 64 is omitted in this specification.

[0021] The in-phase signal filter 65 is a low-pass filter having a predetermined cutoff frequency, and is provided to perform filtering on the output of the in-phase signal mixer 63. The quadrature signal filter 66 is a low-pass filter having the same cutoff frequency as the in-phase signal filter 65, and is provided to perform filtering on the output of the quadrature signal mixer 64. The in-phase signal AD converter 67 is provided to analog-digital convert the output signal of the in-phase signal filter 65 and output the result to the signal processing unit 7. The quadrature signal AD converter 68 is provided to analog-digital convert the output signal of the quadrature signal filter 66 and output the result to the signal processing unit 7.

[0022] The signal processing unit 7 is provided to execute various operations including target detection based on the signal processing result including detection processing of the received signal RX by the receiving circuit unit 6. In this embodiment, the signal processing unit 7 processes the digital signal, i.e., the beat signal, output from the in-phase signal AD converter 67 and the quadrature phase signal AD converter 68 to execute interference suppression processing for suppressing the influence of interference waves included in the received waves. That is, the signal processing unit 7 has a configuration as a digital signal processing unit. Specifically, the signal processing unit 7 may be an on-board microcomputer including a processor 71 and a memory 72. In this case, the processor 71 reads and executes a program from the memory 72 (e.g., a flash memory, etc.), which is a non-transient substantial storage medium, to execute various digital signal processing including interference suppression processing.

[0023] (Operation overview) Below, an overview of the operation of the radar device 1 according to this embodiment and an overview of the execution of a signal processing program executed by a signal processing unit 7 included in the radar device 1 will be described with reference to the drawings, along with the effects achieved by these. In the following description, the radar device 1 according to this embodiment and the signal processing method and signal processing program executed by the signal processing unit 7 may be collectively referred to simply as "this embodiment."

[0024] The transmission signal generating unit 51 generates a transmission signal TX having a predetermined frequency change corresponding to the frequency modulation mode of the transmission wave, and outputs it to the transmission side amplifier 53. The transmission signal TX output from the transmission signal generating unit 51 is distributed by the distributor 52 into a transmission signal TX output to the transmission side amplifier 53 and a local signal LO output to the receiving circuit unit 6. The transmission signal TX input to the transmission side amplifier 53 is power amplified by the transmission side amplifier 53 and output to the transmission antenna 2. As a result, a transmission wave having a predetermined frequency modulation is transmitted from the transmission antenna 2 to the space outside the vehicle.

[0025] When the receiving antenna 3 receives a radar wave (e.g., a desired wave) in a frequency band corresponding to the transmitted wave, the receiving antenna 3 outputs an RF signal. The "desired wave" is a wave reflected from a target of the transmitted wave transmitted from the device itself. The receiving amplifier 61 amplifies the RF signal to generate a received signal RX. The received signal RX is output to each of an in-phase signal mixer 63 and a quadrature-phase signal mixer 64, which are provided in parallel with each other.

[0026] The received signal RX input to the in-phase signal mixer 63 is mixed with the local signal LO in the in-phase signal mixer 63. As a result, an I component, which is the real part of the IF signal, is generated. The I component is filtered by the in-phase signal filter 65, and is analog-to-digital converted by the in-phase signal AD converter 67, and then output to the signal processing unit 7. The received signal RX input to the quadrature signal mixer 64 is mixed with the local signal LO, which has been phase-shifted by the phase shifter 62, in the quadrature signal mixer 64. As a result, a Q component, which is the imaginary part of the IF signal, is generated. The Q component is filtered by the quadrature signal filter 66, and is analog-to-digital converted by the quadrature signal AD converter 68, and then output to the signal processing unit 7. That is, a beat signal, which is a signal obtained by converting an IF signal, which is an analog signal having an I component and a Q component and whose band is limited by a low-pass filter, into a digital signal, is output to the signal processing unit 7.

[0027] When the beat signal is input, the signal processing unit 7 executes various digital signal processing such as FFT on the beat signal. FFT is an abbreviation for fast Fourier transform. This allows the distance to the target and the relative speed to be calculated. The FFT executed by the signal processing unit 7 includes a so-called distance FFT. The distance FFT is an FFT in the time series direction on the beat signal. The time series direction corresponds to the distance direction equivalent to the distance to the target. The result of the distance FFT processing is the frequency spectrum of the beat signal, and the power (i.e., electric power) level for each frequency of the beat signal.

[0028] However, when devices that detect targets by transmitting and receiving radar waves, similar to the radar device 1 according to this embodiment, become widespread, there is a concern about radio wave interference due to interference waves caused by transmission waves from other vehicles. When such radio wave interference occurs, as shown in FIG. 2, the noise floor in the distance FFT spectrum, i.e., the distance map, rises. Then, the peak based on the desired wave is buried in the noise floor, making it difficult to detect the target. In particular, as shown in FIG. 3, the interference wave tends to have a larger voltage amplitude than the desired wave. This is because the voltage amplitude of the interference wave that propagates directly from another device to the own device is inversely proportional to the square of the distance, while the voltage amplitude of the desired wave that is reflected from the own device by the target and received by the own device is inversely proportional to the fourth power of the distance. Therefore, in this embodiment, the signal processing unit 7 performs an interference suppression process that estimates the waveform of the interference wave and subtracts, i.e., removes, the waveform of the interference wave from the waveform of the received wave to suppress the influence of the interference wave contained in the received wave. The waveform estimation process for estimating the waveform of an interference wave and the subtraction process for subtracting the waveform of an interference wave from the waveform of a received wave, both of which are included in the interference suppression process, will be described in detail below.

[0029] FIG. 4 shows the frequency characteristics of the RF and IF signals of the desired wave and the interference wave. LPFis the cutoff frequency of the in-phase signal filter 65 and the quadrature phase signal filter 66, which are low-pass filters. As shown in FIG. 4, the frequency of the interference wave changes linearly with time, and has a different slope from that of the desired wave. Specifically, for example, FIG. 4 shows an example in which the RF signal in the interference wave has a steeper slope than the RF signal in the desired wave. In this case, the IF signal in the desired wave is shown as a horizontal line with a constant frequency. In contrast, the IF signal in the interference wave is shown as a straight line sloping downward to the right, the frequency of which decreases with the passage of time. That is, as shown in FIG. 5, the beat frequency of the interference wave is +F LPF Starts with -F LPF The IF signal in the interference is shown as a sloping straight line ending at +F LPF Starts with -F LPF Since the frequency spectrum has frequency components ending in , the entire floor will rise, as shown in Figure 2.

[0030] 5, the chirp slope K indicates the rate of change over time, i.e., the slope, of the frequency of the IF signal in the interference wave, i.e., the beat frequency. Here, the interference start time Tb is the time when the beat frequency of the interference wave reaches a predetermined value + F LPF On the other hand, the interference end time Te corresponds to the time when the beat frequency of the interference wave reaches a predetermined value -F LPF Then, the beat frequency of the interference wave is expressed by the following formulas (1) and (2). In formula (1), F(t) represents the beat frequency of the interference wave, and t represents the time.

[0031]

number

[0032]

number

[0033]

number

[0034]

number

[0035]

number

[0036]

number

[0037]

number

[0038] y=Dx D=[Φ 1 Φ 2 …Φ M ] When M=N, there is only one solution. On the other hand, when M>N, there are an infinite number of solutions. In general, in the OMP algorithm, to find the solution of the above simultaneous equations, the coefficient vector x, the component x of the coefficient vector x is k First, we calculate the received waveform vector y and each basis vector Φ k Calculate the complex inner product with the received waveform vector y and find the basis vector Φ k , that is, the basis vector Φ that has the highest correlation with the received waveform vector y k Then, find the basis vector Φ k x[1] is calculated from the component x k Only the nonzero component x j (j ≠ k) are all vectors with zero components. That is, x[1] is a vector that contains only one nonzero component and the rest are zero components. Next, x[1] and the found basis vector Φ k Similarly, x[2] can be found from the residual between the product of x and the received waveform vector y, and then x[3], x[4], ... can be found. x[p] is a vector that contains only p non-zero components and the rest are zero components. Hereinafter, the calculation of x[1] will be referred to as the "first step". Therefore, x[p] indicates the coefficient vector x obtained in the pth step of the OMP algorithm. In this way, by repeating the "steps" in the OMP algorithm M times, all components x in the M-row, 1-column coefficient vector x can be found. 1 ,x 2 ,…x M That is, the coefficient vector x can be found as one of an infinite number of solutions.

[0039] However, for large amplitude x kSince the method is carried out from, it is possible to obtain a good approximate solution even if the step is completed in fewer steps than the Mth step. In this respect, in the case of estimating the waveform of an interference wave as in this embodiment, even if the interference wave is expressed as a linear combination of multiple chirplets, there is little physical meaning. This is because an actual interference wave is composed of one chirplet, not multiple chirplets. In addition, during the interference period of the received waveform, the interference wave and the desired wave overlap. Therefore, if the interference wave is expressed as a linear combination of multiple chirplets, the overlapping waveform of the interference wave and the desired wave is modeled, so that in the final subtraction, while the interference wave is removed to a certain extent, the desired wave may also be subtracted.

[0040] Therefore, in this embodiment, the OMP algorithm is first performed as the first step, that is, the first complex inner product calculation and the resulting maximum basis vector Φ k In other words, in the parameter estimation process, the signal processing unit 7 uses the OMP algorithm to calculate the coefficient vector x in the arithmetic equation y=Dx, which represents the received waveform vector y as the product of the dictionary matrix D and the coefficient vector x, so that the coefficient vector x contains only one non-zero component and the rest are zero components. Then, the waveform estimation process is performed by selecting the coefficient vector x containing only one non-zero component and one basis vector Φ k The waveform of the interference wave is estimated by this.

[0041] Specifically, in the parameter estimation process, the signal processing unit 7 selects a basis vector Φ that maximizes the inner product with the received waveform vector y from the dictionary matrix D. k Then, select only one basis vector Φ k Based on the received waveform vector y, the coefficient vector x, i.e., x[1], is calculated. k and coefficient vector x include the estimation results of the above four parameters. Then, signal processing unit 7 estimates the waveform of the interference wave based on the estimation results of the four parameters, and subtracts the estimated waveform of the interference wave from the waveform of the received wave.

[0042] More specifically, in the parameter estimation process, the signal processing unit 7 selects a basis vector Φ k The process of selecting only one basis vector Φ k The process of updating the dictionary matrix D based on the above is repeated a predetermined number of times. k The dictionary matrix D, which is a set of the above, is updated so that the range of the chirp slope K is smaller after the update than before the update. Specifically, the range of the chirp slope K is set to be smaller than the range of the selected basis vector Φ k Then, the signal processing unit 7 updates the dictionary matrix D updated a predetermined number of times to a value in the vicinity of the value corresponding to the basis vector Φ k If you select only one, the selected basis vector Φ k and the received waveform vector y, a coefficient vector x is calculated.

[0043] Thus, in this embodiment, after executing the first OMP algorithm, the signal processing unit 7 updates the dictionary matrix D for the first time, and executes the second OMP algorithm using the updated dictionary matrix D. After executing the second OMP algorithm, the signal processing unit 7 updates the dictionary matrix D for the second time, and executes the third OMP algorithm using the updated dictionary matrix D. After executing the Pth OMP algorithm, the signal processing unit 7 obtains the coefficient vector x=x[1], and uses this to estimate the waveform of the interference wave. As is clear from the above, the number of "steps" in each OMP algorithm is one.

[0044] Here, the number of times the dictionary matrix D is updated, i.e., the number of times the OMP algorithm is executed, is set in advance according to the assumed conditions of use of the radar device 1. If the number of times the OMP algorithm is executed is P, this value P is the number of times the dictionary matrix D is updated plus "1". Specifically, for example, the number of times P at which the noise floor level does not improve even after the (P+1)th execution can be obtained in advance by experiment or computer simulation. This is because, just as the frequency bin number in the distance FFT is limited and it is not possible to guess the frequency of the desired wave, it is theoretically impossible to guess the frequency of the interference wave and perform complete interference removal. Therefore, there is no point in making the number of times the OMP algorithm is executed unlimited.

[0045] However, if the amount of calculation in each OMP algorithm is large, it is possible that P+1 calculations will not be completed during one chirp, and the interference removal process will not be completed. Therefore, it is preferable to limit the number of times the OMP algorithm is executed (for example, to two or less). Therefore, in this embodiment, the signal processing unit 7 calculates the basis vector Φ k The initial range of the chirp slope K and the interference onset time Tb is not adopted over the entire possible range. Instead, estimates of the chirp slope K and the interference onset time Tb are calculated before the first execution of the OMP algorithm, and the basis vector Φ k The range of the chirp slope K and the interference start time Tb is narrowed down to a narrower range that includes the approximate values.

[0046] Specifically, before the first execution of the OMP algorithm, as preprocessing, the interference start time Tb and the interference end time Te are estimated by threshold testing, and the first chirp slope K is roughly estimated from the interference start time Tb and the interference end time Te. K=2F LPF / (Te-Tb) In the above formula, K is an approximate value of the chirp slope K, Tb is an approximate value of the interference start time Tb, and Te is an approximate value of the interference end time Te.

[0047] The method of calculating the approximate values ​​of the interference start time Tb and the interference end time Te may be, for example, as shown in FIG. 6, a time point when the absolute value calculated from the real part and the imaginary part of the beat signal crosses a predetermined threshold value TH as the interference start time Tb and the interference end time Te. Specifically, the timing at which the absolute value changes from a state smaller than the threshold value TH to a state larger than the threshold value TH may be specified as the approximate value of the interference start time Tb. Then, the timing at which the absolute value changes from a state larger than the threshold value TH to a state smaller than the threshold value TH may be specified as the approximate value of the interference end time Te. In FIG. 6, the real part, the imaginary part, and the absolute value of the beat signal are shown from the top to the bottom with time. Note that this threshold value TH may be a fixed value or may be an amount that varies according to a predetermined condition. For example, the threshold value TH may be the same or different when specifying the approximate value of the interference start time Tb and when specifying the approximate value of the interference end time Te.

[0048] Alternatively, the time when the absolute value waveform has the first step over time may be specified as the interference start time Tb, and the time when the absolute value waveform has the last step over time may be specified as the interference end time Te, where the step refers to the gradient of the change in absolute value over time that is equal to or greater than a predetermined reference value.

[0049] Fig. 7 shows a first comparative example in which the OMP algorithm is performed multiple times, specifically 15 times, to obtain a coefficient vector x=x

[15] . In the first comparative example shown in Fig. 7, 20 discrete chirp slopes K are prepared. In addition, a total of 300 interference start times Tb, from 1 to 300, are prepared.

[0050] 8 to 13 show an additional comparative example in which the number of executions of the OMP algorithm is set to the above-mentioned value P=5. In the additional comparative example, after the OMP algorithm is once terminated in the first step, the update of the dictionary matrix D and the first step of the OMP algorithm using the updated dictionary matrix D are repeated four times to select the basis vector Φ kIn this additional comparative example, 17 discrete chirp slopes K are prepared. A total of 512 interference start times Tb are prepared, ranging from 1 to 512. However, in this additional comparative example, the chirp slope K and the interference start time Tb are not estimated before the first execution of the OMP algorithm, and the basis vector Φ k The range of the chirp slope K and the interference start time Tb is gradually narrowed from all possible ranges.

[0051] FIG. 14 shows how the value of the chirp slope K converges in an additional comparative example. That is, FIG. 14 shows how the value of the chirp slope K converges in FIG. 8 to FIG. 12 or FIG. 13. In the figure, the current value, i.e., the value selected by the current OMP algorithm, is surrounded by a solid-line square, and the previous value is surrounded by a dashed-line square. In FIG. 14, the discrete values ​​of the chirp slope K are set geometrically. Specifically, for example, the chirp slope K in the first OMP is set so that adjacent values ​​are 1.3200 times larger when arranged in ascending order of value. Similarly, the chirp slope K in the second OMP is set so that adjacent values ​​are 1.0533 times larger. The chirp slope K in the third OMP is set so that adjacent values ​​are 1.0089 times larger. The chirp slope K in the fourth OMP is set so that adjacent values ​​are 1.0015 times larger. The chirp slope K in the fifth OMP is set so that adjacent values ​​are 1.0002 times larger. Figure 15 shows how the value of the interference start time Tb converges in an additional comparative example. In Figure 15, the current value, i.e., the value selected by the current OMP algorithm, is enclosed in a solid-line square.

[0052] 7 to 13, the matrix expressions and waveforms are schematic and simplified for convenience of illustration, from the viewpoint of helping to understand the effect of the additional comparative example. That is, for example, the received waveform vector y is actually a complex waveform, but for simplicity of illustration, it is shown schematically as if it were a real waveform. The number of sampling points shown in the schematic waveform of the received waveform vector y should also be large (for example, 512, which is a power of 2), but to avoid complicating the illustration, most of them are omitted and only a portion is shown schematic. Also, in FIG. 7, "K=54.293e12" represents K=54.293×10 12 The same is true for "K=84e12" and the like. The same is true for Figs. 8 to 13. In Fig. 14, "×10 12 " is omitted. Also, for the sake of simplicity, "K=84e12" in FIG. 8 should be written as "K=84.8e12" according to the table in FIG. 14, but considering that this is a stage before the values ​​converge, the part after the decimal point is omitted. The same is true for FIG. 9 and FIG. 10.

[0053] In the first comparative example shown in Fig. 7, a single dictionary matrix D is used without updating, and the OMP algorithm is executed for multiple steps to calculate a coefficient vector x including multiple non-zero components. However, according to the first comparative example, the noise floor is not lowered sufficiently, and the signal-to-noise ratio, which is one of the indicators of interference removal, remains low (for example, about 10 dB). This is because, as described above, an interference wave that should be a single chirplet is expressed as a linear sum of multiple chirplets, and the interference wave is not correctly represented physically or mathematically, and as a result, the interference cannot be completely removed.

[0054] In contrast, in the additional comparative example, the basis vector Φ k Then, (K, Tb) = (48.7 × 10 12 ,124) is selected. Then, the basis vector Φ kThe chirp slope K value at 48.7×10 12 As shown in FIG. 9, the dictionary matrix D is updated using the above. Here, in the dictionary matrix D used in the first OMP algorithm, the range of values ​​of the chirp slope K is set wide. On the other hand, in the updated dictionary matrix D, the range of values ​​of the chirp slope K is set wide depending on the selected basis vector Φ k The value at 48.7×10 12 is set as the central value and is set narrower than the previous time.

[0055] Next, the second OMP algorithm is used to find the basis vector Φ k Then, (K, Tb) = (54.0 × 10 12 ,123) is selected. Then, the basis vector Φ k The chirp slope K value at 54.0×10 12 The dictionary matrix D is updated as shown in FIG. 10 using the above. In the updated dictionary matrix D, the range of values ​​of the chirp slope K is determined based on the selected basis vector Φ k Value at 54.0×10 12 is set as the central value and is set narrower than the previous time.

[0056] Next, the third OMP algorithm is used to find the basis vector Φ k Then, (K, Tb) = (54.48 × 10 12 ,123) is selected. Then, the basis vector Φ k The chirp slope K value at 54.48×10 12 The dictionary matrix D is updated as shown in FIG. 11 using the above. In the updated dictionary matrix D, the range of values ​​of the chirp slope K is determined based on the selected basis vector Φ k The value at 54.48×10 12 is set as the central value and is set narrower than the previous time.

[0057] Similarly, the fourth OMP algorithm produces the basis vector Φ k Then, (K, Tb) = (54.32 × 10 12 ,122) is selected. Then, the basis vector Φ k The chirp slope K value at 54.32×10 12 The dictionary matrix D is updated as shown in FIG. 12 using the above. In the updated dictionary matrix D, the range of values ​​of the chirp slope K is determined based on the selected basis vector Φ k The value at 54.32×10 12 is set as the central value and is set narrower than the previous time.

[0058] Then, the fifth OMP algorithm is used to find the basis vector Φ k Then, (K, Tb) = (54.293 × 10 12 , 122) is selected. Then, such a basis vector Φ k The waveform of the interference wave is estimated based on the coefficient vector x calculated using the current dictionary matrix D. k Then, (K, Tb) = (54.293 × 10 12 ,122) is selected. The selected basis vector Φ k The waveform of the interference wave differs only in amplitude A and phase θ.

[0059] FIG. 16A and FIG. 16B show how the influence of the interference wave is reduced from the waveform of the received wave by the additional comparative example. In the figure, (0) shows the waveform before the interference suppression process, (1) shows the effect of reducing the influence of the interference wave by the first OMP, (2) shows the effect of reducing the influence of the interference wave by the second OMP, and (3) shows the effect of reducing the influence of the interference wave by the third OMP. Also, (4) shows the effect of reducing the influence of the interference wave by the fourth OMP, and (5) shows the effect of reducing the influence of the interference wave by the fifth OMP. In FIG. 16A and FIG. 16B, the vertical axis indicates voltage, and the horizontal axis indicates sampling points corresponding to time. The influence of the interference wave is almost completely eliminated by the fourth OMP, and the influence of the interference wave is almost completely eliminated by the fifth OMP. Note that, for the sake of clarity, FIG. 16A and FIG. 16B show the reduction effect of the interference wave subtracted at each of the first to fifth OMPs, and in the actual additional comparative example, the subtraction of the interference wave is only performed after the fifth OMP.

[0060] 17 and 18 show, in the same format as in FIGS. 14 and 15, how the values ​​of the chirp slope K and the interference start time Tb converge according to this embodiment. One of the differences between this embodiment and the additional comparative example is that the basis vector Φ k The difference is whether the range of the chirp slope K and the interference start time Tb are limited based on the estimated value or not. The former is limited, and the latter is not limited. Furthermore, the difference between this embodiment and the additional comparative example is the number of times the OMP is executed. The other processing contents are the same as the above-mentioned processing of the additional comparative example.

[0061] In the example in Figure 17, the approximate chirp slope K calculated in the preprocessing is 54.48 × 10 12 The basis vector Φ used for the dictionary matrix D at the beginning of the OMP algorithm is k The range of the chirp slope K is approximately ±0.64×10 12 In other words, the basis vector Φ used for the dictionary matrix D at the beginning of the OMP algorithm is limited to the range kThe chirp slope K of is limited to a range of less than 3% of the estimated value including the estimated value, or may be limited to a range of less than 5% of the estimated value including the estimated value, or may be limited to a range of less than 10% of the estimated value including the estimated value, or may be limited to a range of less than 20% of the estimated value including the estimated value.

[0062] 17, the discrete values ​​of the chirp slope K are set geometrically. Specifically, for example, the chirp slope K in the first OMP is set so that adjacent values ​​are 1.0015 times larger. The chirp slope K in the second OMP is set so that adjacent values ​​are 1.0002 times larger.

[0063] In the example of FIG. 18, the estimated value of the interference start time Tb calculated in the preprocessing is 122, and the basis vector Φ k The range of the interference start time Tb is limited to a range of ±3 around the estimated value. That is, the basis vector Φ k The interference start time Tb is limited to a range of less than 5% of the estimated value including the estimated value, or may be limited to a range of less than 10% of the estimated value including the estimated value, or may be limited to a range of less than 20% of the estimated value.

[0064] Figure 19 shows how the influence of the interference wave is reduced from the waveform of the received wave according to this embodiment. In the figure, (0) shows the waveform before the interference suppression process, (1) shows the reduction effect of the influence of the interference wave by the first OMP, and (2) shows the reduction effect of the influence of the interference wave by the second OMP. In Figure 19, the vertical axis represents voltage, and the horizontal axis represents sampling points corresponding to time. As shown in this figure, by one or two OMPs, the influence of the interference wave is almost removed and the matching is completed. That is, compared with the additional comparative examples in FIGS. 16A and 16B, the number of times of OMP required to sufficiently remove the influence of the interference wave can be reduced. Note that Figure 19 shows the reduction effect by subtracting the interference wave at each of the first and second OMPs for the sake of easy understanding, and in the actual example, the subtraction of the interference wave is performed only after the second OMP.

[0065] In this way, the additional comparative example and this embodiment calculate the coefficient vector x containing only one non-zero component by the one-step OMP algorithm, and estimate the waveform of the interference wave by such a coefficient vector x and one basis vector Φ k Specifically, this embodiment repeatedly performs the one-step OMP algorithm a small number of times while updating the dictionary matrix D. As a result, it is possible to improve the estimation accuracy of the waveform of the interference wave while reducing the computational load. Therefore, according to the additional comparative example and this embodiment, in an environment where radio interference occurs, it is possible to reduce the influence of the interference wave in the received signal RX even better than before.

[0066] Also, in the additional comparative example and this embodiment, the values of a plurality of chirp slopes K in the dictionary matrix D are set discretely and geometrically. The ratio is decreased for each number of OMPs, that is, for each update of the dictionary matrix D. The dictionary matrix D is generated by setting each of a plurality of basis vectors Φ k by combining the discretely set values of M1 chirp slopes K and the values of M2 (where M1 < M2 and M1 × M2 = M) interference start times Tb. That is, in the dictionary matrix D, the basis vector Φ kThe interference start time Tb is set using all possible sampling times in order to ensure accurate estimation of the waveform of the interference wave.

[0067] Fig. 20 shows distance maps in this embodiment, comparing those before and after interference removal, i.e., after the influence of the interference wave is almost completely removed by the second OMP. In Fig. 20, the horizontal axis indicates frequency bins corresponding to distance, and the vertical axis is in dB. The single sharp peak on the left side is the peak of the desired wave. As shown in Fig. 20(A), before interference removal, the noise floor is raised by the interference wave, resulting in a low S / N ratio. In contrast, as shown in Fig. 20(B), after interference removal, the S / N ratio is improved to about 40 dB.

[0068] 21 is a flowchart showing the flow of a series of processes in this embodiment. The series of processes is realized by the processor 71 reading necessary data such as programs and initial values ​​from the memory 72 and executing the programs.

[0069] First, in S1801, the processor 71 calculates an approximate value of the interference start time Tb and an approximate value of the interference end time Te from the received waveform by threshold testing, as described with reference to Fig. 6. Next, in S1801, the processor 71 calculates an approximate value of the chirp slope K based on the estimated interference start time Tb and interference end time Te. K=2F LPF / (Te-Tb) It is calculated as follows.

[0070] Next, in S1802, the processor 71 sets the ranges of the interference start time Tb and the chirp slope K to be used in the dictionary matrix D based on the estimated values ​​of the interference start time Tb and the chirp slope K obtained in steps S1800 and S1801. Specifically, it sets a range that includes the interference start time Tb estimated in step S1800 and has the interference start time Tb as its center, and a range that includes the chirp slope K calculated in S1802 and has the chirp slope K as its center. Then, based on the determined ranges, it calculates the basis vector Φ k In this way, an initialized dictionary matrix D is generated. Next, in S1803, the processor 71 initializes the received waveform vector y and each basis vector Φ k Then, in S1804, the processor 71 calculates the basis vector Φ k Then, in S1805, the processor 71 selects the basis vector Φ k It is determined whether the selection is the Pth time.

[0071] If it is less than the Pth time (i.e., S1805=NO), the processor 71 executes the process of S1806, and then returns the process to S1803. In S1806, the processor 71 calculates the basis vector Φ k That is, the processor 71 updates the dictionary matrix D. On the other hand, if it is the Pth time (that is, S1805=YES), the processor 71 advances the process to S1807 and subsequent steps.

[0072] In S1807, the processor 71 obtains a coefficient vector x, i.e., x[1]. In S1808, the processor 71 obtains the waveform of the interference wave. In S1809, the processor 71 subtracts the waveform of the interference wave obtained in S1808 from the waveform of the received wave. In S1810, the processor 71 performs distance FFT on the waveform after subtraction. After step S1810, the interference suppression process ends.

[0073] In the above processing, the processes of steps S1800 to S1809 correspond to interference suppression processing. Moreover, in the interference suppression processing, the processes of steps S1800 to S1808 correspond to waveform estimation processing, and the process of step S1809 corresponds to subtraction processing. Moreover, in the waveform estimation processing, the processes of steps S1800 to S1807 correspond to parameter estimation processing, and the process of step S1808 corresponds to interference waveform calculation processing. Moreover, in the parameter estimation processing, the processes of steps S1802 to S1806 correspond to OMP processing, and the processes of steps S1800 and S1801 correspond to pre-processing.

[0074] [1] As described above, the signal processing unit 7 mixes the received wave, which is the radar wave received by the receiving antenna 3, with the local signal LO to generate a predetermined cutoff frequency F LPS A beat signal is obtained by filtering the received signal with the above filter. Then, based on the result of the signal processing on the beat signal, an interference suppression process is performed to suppress the influence of an interference wave contained in the received wave.

[0075] In the interference suppression process, a waveform estimation process is performed to estimate the waveform of the interference wave, and a subtraction process is performed to subtract the estimated waveform of the interference wave from the waveform of the received wave.

[0076] Furthermore, in the waveform estimation process, a parameter estimation process is performed to estimate parameters such as the interference start time Tb, the chirp slope K, the amplitude A, and the phase θ, and an interference waveform calculation process is performed to calculate the waveform of the interference wave based on the estimated parameters.

[0077] Furthermore, in the parameter estimation process, an OMP process is executed to calculate the coefficient vector x using an orthogonal matching pursuit algorithm so that the coefficient vector x contains only one non-zero component and the rest are zero components.

[0078] In addition, in the parameter estimation process, a pre-processing is performed, that is, an approximate value of the chirp slope K is calculated before the OMP process, and the range of the value of the chirp slope K used in the basis vectors of the dictionary matrix D is restricted so that it falls within a predetermined range including the calculated approximate value.

[0079] In this way, the radar device 1 uses an orthogonal matching pursuit algorithm to calculate the coefficient vector x in the calculation formula y=Dx so that it contains only one non-zero component and the rest are zero components. This reduces the possibility of removing not only the interference wave but also the desired wave.

[0080] Furthermore, the radar device 1 calculates an approximate value of the chirp slope K before the OMP process, and based on the approximate value, limits the value of the chirp slope K used in the basis vectors of the dictionary matrix D. In this way, it becomes possible to effectively remove interference waves with fewer OMP processes compared to a case in which basis vectors are prepared for the entire theoretically possible range of the chirp slope K.

[0081] [2] In addition, in the pre-processing, the signal processing unit 7 performs a threshold test on the beat signal to identify the interference start time Tb and the interference end time Te. The identified interference start time Tb, interference end time Te, and cutoff frequency F LPS In this way, an approximate value of the chirp slope K can be calculated by a simple method of threshold testing for the beat signal.

[0082] [3] More specifically, the signal processor 7 determines the timing at which the absolute value of the beat signal changes from being smaller than a predetermined threshold value TH to being larger than it, as the interference start time Tb. The signal processor 7 determines the timing at which the absolute value of the beat signal changes from being larger than the threshold value TH to being smaller than it, as the interference end time Te. In this way, the absolute value of the beat signal in the interference wave is sufficiently larger than that of the desired wave, making it possible to calculate an approximate value of the chirp slope K.

[0083] [4] More specifically, the chirp slope K can be estimated as K=2F LPF / (Te-Tb) Therefore, the chirp slope K can be roughly calculated using a simple arithmetic formula.

[0084] Second embodiment Next, a second embodiment will be described. This embodiment is different from the first embodiment in the method of estimating the interference start time Tb and the chirp slope K before the execution of OMP. Specifically, in this embodiment, the processor 71 performs the process shown in FIG. 22 instead of the process shown in FIG. 21.

[0085] Specifically, first, in step S110, the processor 71 performs a short-time Fourier transform (STFT) on the acquired beat signal. STFT is an abbreviation for Short-Time Fourier Transform.

[0086] By performing the short-time Fourier transform, a map such as that shown in FIG. 23 is obtained. This map is a two-dimensional map (i.e., an STFT graph) showing the power distribution of the beat signal for each time and frequency during one chirp period. In this two-dimensional map, a bin is provided for each time interval that is one unit of the short-time Fourier transform on the horizontal axis indicating time, and a bin is provided for each frequency interval that is one unit of the short-time Fourier transform on the vertical axis indicating space. Therefore, the two-dimensional map is a matrix of a combination of multiple columns of time bins and multiple rows of space bins. The brightness of each cell in this matrix indicates the power (i.e., signal strength) of the beat signal at the time and frequency corresponding to that cell. The higher the brightness of the cell (i.e., the stronger the degree of whiteness), the higher the power of the beat signal in that cell.

[0087] Subsequently, the processor 71 performs the first type pre-processing in step S120. In the first type pre-processing, as shown in Fig. 24, first, in step S121, all cells included in the row MR1 of the minimum frequency in the two-dimensional map are selected as targets.

[0088] Next, in step S122, the maximum peak power among the powers of all cells in the target frequency row is identified, and it is determined whether the identified maximum peak power is equal to or greater than a predetermined judgment value. If it is equal to or greater than the judgment value, the process proceeds to step S123. If it is less than the judgment value, it is determined that the first type pre-processing did not find an interference wave, and the first type pre-processing is terminated without calculating the approximate value of the chirp slope K and the approximate value of the interference start time Tb. Note that the judgment value is preset to be higher than the power assumed to be the power of the desired wave in the beat signal, so as to detect the interference wave and not the desired wave.

[0089] In step S123, the time and frequency of cell MX1 with the maximum peak power obtained in the previous step S122 are recorded. Then, in step S124, all cells included in the next time sequence MC1 are targeted. The "next time sequence" refers to the time sequence immediately following the time of cell MX1 with the maximum peak power.

[0090] Next, in step S125, the maximum peak power of all the cells in the target time sequence MC1 is identified, and it is determined whether the identified maximum peak power is equal to or greater than a predetermined threshold. If it is equal to or greater than the threshold, the process returns to step S123, records the time and frequency of cell MX2 with the maximum peak power obtained in the previous step S125, and then executes steps S124 and S125. If it is less than the threshold, it is determined that the end of the interference wave has been reached, and the process proceeds to step S126.

[0091] In this way, the processor 71 repeats the processes of steps S123, S124, and S125 many times to detect and record the positions of cells in the two-dimensional map where interference waves exist (i.e., cells with the maximum peak power in each row) over time. In the example of Fig. 23, the processes of steps S123, S124, and S125 are repeated while moving the time row one by one to the right in the figure (i.e., along the time course) until the next time row becomes time row MC2.

[0092] When the processor 71 determines that the end of the interference wave has been reached, it ends the repetition of steps S123, S124, and S125 and proceeds to step S126. In step S126, the processor 71 calculates an approximate value of the chirp slope K and an approximate value of the interference start time Tb for the interference wave using the times and frequencies of the multiple cells recorded in step S123.

[0093] For example, the equation of the line ML1 of the interference wave that approximates the time and frequency of multiple cells in the two-dimensional map is estimated by the least squares method. Then, the chirp slope K and the interference start time Tb according to the equation may be set as approximate values, respectively. Alternatively, an estimation method other than the least squares method may be used.

[0094] After step S126, the first type pre-processing is completed. By performing such first type pre-processing, it is possible to calculate an approximate value of the chirp slope K of the interference wave extending from the row of the minimum frequency in the two-dimensional map to the upper right and the interference start time Tb.

[0095] After the first type pre-processing in step S120 is completed, the processor 71 proceeds to step S130 to determine whether the estimated values ​​of the chirp slope K and the interference start time Tb have been calculated in the first type pre-processing, and if so, proceeds to step S180, and if not, proceeds to step S140.

[0096] In step S140, a second type pre-processing is performed. The second type pre-processing is a process for calculating an approximate value of the chirp slope K of the interference wave extending from the row of the maximum frequency in the two-dimensional map to the lower right, and an approximate value of the interference start time Tb. In the second type pre-processing, as shown in Fig. 25, first, in step S141, all cells included in the row of the maximum frequency in the above two-dimensional map are selected as targets.

[0097] Next, in step S142, as in step S122 of the first type pre-processing, it is determined whether the maximum peak power among the powers of all cells in the target frequency row is equal to or greater than a predetermined judgment value. If it is equal to or greater than the judgment value, the process proceeds to step S143. If it is less than the judgment value, it is determined that no interference wave was found in the second type pre-processing, and the second type pre-processing is terminated without calculating the approximate value of the chirp slope K and the approximate value of the interference start time Tb.

[0098] The process contents of steps S143, S144, and S145 are the same as steps S123, S124, and S125 of the first pre-processing. That is, in step S143, the time and frequency of the cell with the maximum peak power obtained in the immediately preceding step S142 are recorded. Then, in step S144, all cells included in the next time sequence are targeted. Then, in step S145, it is determined whether the maximum peak power among the powers of all cells in the targeted time sequence is equal to or greater than a predetermined judgment value. If it is equal to or greater than the judgment value, the process proceeds to step S143, and the time and frequency of the cell with the maximum peak power obtained in the immediately preceding step S145 are recorded. If it is less than the judgment value, it is determined that the end of the interference wave has been reached, and the process proceeds to step S146.

[0099] In this manner, the processor 71 repeats the processes of steps S143, S144, and S145 to detect and record the positions of cells in which interference waves exist in the two-dimensional map over time.

[0100] When the processor 71 determines that the end of the interference wave has been reached, in step S146, it calculates an approximate value of the chirp slope K and an approximate value of the interference start time Tb for the interference wave using the times and frequencies of the multiple cells recorded in step S143 up to this point. The calculation method may be the same as or different from that in step S126 of the first type pre-processing.

[0101] After step S146, the second type pre-processing is completed. By performing such second type pre-processing, it is possible to calculate an approximate value of the chirp slope K of the interference wave extending from the row of the maximum frequency in the two-dimensional map to the lower right and an approximate value of the interference start time Tb.

[0102] After the second type pre-processing in step S140 is completed, the processor 71 proceeds to step S150 to determine whether the chirp slope K and the interference start time Tb have been calculated in the second type pre-processing, and if so, proceeds to step S180, and if not, proceeds to step S160.

[0103] In step S160, a third type pre-processing is performed. The third type pre-processing is a process for calculating an approximate value of the chirp slope K of the interference wave extending from the first time column in the two-dimensional map to the upper right or lower right, and an approximate value of the interference start time Tb. In the third type pre-processing, as shown in Fig. 26, first, in step S161, all cells included in the first time column of the above two-dimensional map are selected as targets.

[0104] Next, in step S162, the maximum peak power among the powers of all cells in the target time sequence is identified, and it is determined whether the identified maximum peak power is equal to or greater than a predetermined judgment value. If it is equal to or greater than the judgment value, the process proceeds to step S163. If it is less than the judgment value, it is determined that no interference wave was found in the third type pre-processing, and the third type pre-processing is terminated without calculating the approximate value of the chirp slope K and the approximate value of the interference start time Tb.

[0105] The process contents of steps S163, S164, and S165 are the same as steps S123, S124, and S125 of the first pre-processing. That is, in step S163, the time and frequency of the cell with the maximum peak power obtained in the immediately preceding step S162 are recorded. Then, in step S164, all cells included in the next time sequence are targeted. Then, in step S165, it is determined whether the maximum peak power among the powers of all cells in the targeted time sequence is equal to or greater than a predetermined judgment value. If it is equal to or greater than the judgment value, the process proceeds to step S163, and the time and frequency of the cell with the maximum peak power obtained in the immediately preceding step S165 are recorded. If it is less than the judgment value, it is considered that the end of the interference wave has been reached, and the process proceeds to step S166.

[0106] In this manner, the processor 71 repeats the processes of steps S163, S164, and S165 to detect and record the positions of cells in which interference waves exist in the two-dimensional map over time.

[0107] When the processor 71 determines that the end of the interference wave has been reached, in step S166, it calculates an approximate value of the chirp slope K and an approximate value of the interference start time Tb for the interference wave using the times and frequencies of the multiple cells recorded in step S163. The calculation method may be the same as or different from that in step S126 of the first type pre-processing.

[0108] After step S166, the third type pre-processing is completed. By performing such third type pre-processing, it is possible to calculate an approximate value of the chirp slope K of the interference wave extending from the first time column in the two-dimensional map to the upper right or lower right, and the interference start time Tb.

[0109] After the third type pre-processing in step S160 is completed, processor 71 proceeds to step S170 to determine whether or not an approximate value of chirp slope K and an approximate value of interference start time Tb have been calculated in the third type pre-processing. If they have been calculated, the process proceeds to step S180, and if they have not been calculated, the process proceeds to step S190, assuming that no interference wave exists in the beat signal. If the process proceeds from step S160 to step S190, a distance FFT is performed on the beat signal in the same manner as in step S1810 in FIG. 21, without performing interference removal.

[0110] In step S180, interference removal processing is performed based on the estimated value of chirp slope K and interference start time Tb calculated before step S170. The interference removal processing is similar to the processing from step S1802 to step S1809 in Fig. 21. After step S180, the process proceeds to step S190. When proceeding from step S180 to step S190, distance FFT is performed on the beat signal from which interference has been removed, similar to step S1810 in Fig. 21. In this embodiment, the processing from steps S110 to S170 corresponds to pre-processing.

[0111] Figure 27 shows how the influence of interference waves is reduced from the waveform of a received wave according to this embodiment. In the figure, (0) shows the waveform before interference suppression processing, (1) shows the effect of reducing the influence of interference waves by the first OMP, and (2) shows the effect of reducing the influence of interference waves by the second OMP. In Figure 27, the vertical axis indicates voltage, and the horizontal axis indicates sampling points corresponding to time.

[0112] As shown in this figure, the influence of the interference wave is almost completely eliminated by one or two OMPs, and matching is completed. That is, the number of OMPs required to sufficiently eliminate the influence of the interference wave can be reduced, compared with the additional comparative example of Fig. 16A and Fig. 16B of the first embodiment. Note that Fig. 27 shows the reduction effect of subtracting the interference wave after OMP1 and OMP2 for the sake of clarity, and in an actual embodiment, subtraction of the interference wave is performed only after the second OMP.

[0113] [1] As described above, in preprocessing, the signal processing unit 7 performs a short-time Fourier transform on the beat signal to calculate the power distribution of the beat signal for each time and frequency, and calculates an approximate value of the chirp slope K based on the calculated power distribution. In this way, it is possible to calculate an approximate value of the chirp slope K with a more detailed calculation than in the first embodiment.

[0114] [2] The signal processing unit also performs a process of recording the time and frequency pair at which the maximum power is achieved in the row of the minimum frequency in the two-dimensional map. Then, the signal processing unit repeats the process of recording the time and frequency pair at which the maximum power is achieved in the next time column of the two-dimensional map. Then, based on the recorded time and frequency pairs, an approximate value of the chirp slope K is calculated. In this way, an approximate value of the chirp slope K of the interference wave extending from the row of the minimum frequency to the upper right in the two-dimensional map can be calculated.

[0115] [3] The signal processor also performs a process of recording the time and frequency pair at which the maximum power is realized in the maximum frequency row of the two-dimensional map. Then, the signal processor repeats the process of recording the time and frequency pair at which the maximum power is realized in the next time column of the two-dimensional map. Then, based on the recorded time and frequency pairs, an approximate value of the chirp slope K is calculated. In this way, an approximate value of the chirp slope K of the interference wave extending from the maximum frequency row to the lower right in the two-dimensional map can be calculated.

[0116] [4] The signal processor also records the time and frequency pair at which the maximum power is achieved in the first time column of the maximum frequency row of the two-dimensional map. It then repeats the process of recording the time and frequency pair at which the maximum power is achieved in the next time column after the last recorded time in the two-dimensional map. It then calculates an estimate of the chirp slope K based on the recorded time and frequency pairs.

[0117] In this way, it is possible to calculate an approximate value of the chirp slope K of the interference wave extending to the upper right or lower right from the first time series in the two-dimensional map. In other words, it is possible to calculate an approximate value of the chirp slope K of the interference wave at the chirp end.

[0118] It should be noted that the present embodiment and the first embodiment have similar configurations and operations, and therefore similar effects can be obtained.

[0119] Third embodiment Next, a third embodiment will be described. This embodiment differs from the second embodiment in the processing contents of the signal processing unit 7. Specifically, the processor 71 of this embodiment executes the processing of FIG. 28 instead of the processing of FIG. 22. In the processing of FIG. 28, the processing of step S195 is added to the processing of FIG. 22, and the processing of steps S110 to S190 is repeated, so that a plurality of interference waves can be removed. The processing contents of steps S110 to S190 are the same as those of FIG. 22.

[0120] After performing the distance FFT process in step S190, the processor 71 proceeds to step S195 to determine whether or not the noise floor value in the distance FFT spectrum, i.e., the distance map, is equal to or less than a specified value. If the noise floor value is equal to or less than the specified value, it is assumed that all interference waves have been removed, and the process of FIG. 28 ends. If the noise floor value is greater than the specified value, the process returns to step S110 to perform a short-time Fourier transform on the beat signal. However, the beat signal at this time is the beat signal after interference removal in the immediately preceding step S180.

[0121] By performing such processing by the processor 71, in the two-dimensional map obtained by the first short-time Fourier transform, as shown in FIG. 29, there are two interference waves CW1 and CW2, whereas in the two-dimensional map obtained by the second short-time Fourier transform, the interference waves are reduced. That is, as shown in FIG. 30, one of the two interference waves CW1 and CW2 is removed, and one interference wave CW2 remains. Then, all interference waves are removed from the beat signal by the processing of steps S120 to S180 after the second short-time Fourier transform. As a result, if the beat signal from which the two interference waves have been removed is subjected to a short-time Fourier transform, a two-dimensional map from which the interference waves have been removed is obtained as shown in FIG. 31. It is determined in step S195 that the noise floor of the distance FFT spectrum obtained by the distance FFT for the beat signal from which the two interference waves have been removed is equal to or less than a specified value. As shown in FIG. 29, FIG. 30, and FIG. 31, as the interference waves are removed, the desired TW becomes more noticeable.

[0122] Fig. 32 shows how the influence of multiple interference waves is reduced from the waveform of a received wave according to this embodiment. In the figure, (0) shows the waveform before interference suppression processing, (11) shows the effect of reducing the influence of interference waves by the first OMP in the first round, and (12) shows the effect of reducing the influence of interference waves by the second OMP in the first round. Also, (21) shows the effect of reducing the influence of interference waves by the first OMP in the second round, and (22) shows the effect of reducing the influence of interference waves by the second OMP in the second round. Here, the first round and the second round refer to the number of loops of steps S110 to S195. In Fig. 32, the vertical axis indicates voltage, and the horizontal axis indicates sampling points corresponding to time.

[0123] As shown in this figure, the influence of the interference wave is almost completely eliminated and matching is completed by performing OMP once or twice in each cycle. In other words, the number of OMPs required to sufficiently eliminate the influence of the interference wave can be reduced compared to the comparative examples of Figures 33A and 33B of the first embodiment.

[0124] 33A and 33B show how the influence of interference waves is reduced from the waveform of a received wave in a comparative example to this embodiment. In the figures, (0) shows the waveform before interference suppression processing, (11) to (15) show the effect of reducing the influence of interference waves by the first to fifth OMPs in the first round, and (21) to (25) show the effect of reducing the influence of interference waves by the first to fifth OMPs in the second round. The difference between this embodiment and this comparative example is the basis vector Φ k The difference is whether the range of the chirp slope K and the interference start time Tb are limited based on the estimated value or not. The former is limited, and the latter is not limited. Furthermore, the difference between this embodiment and the comparative example is the number of times the OMP is executed.

[0125] 33A and 33B, the vertical axis indicates voltage, and the horizontal axis indicates sampling points corresponding to time. The fourth OMP of each cycle largely removes the influence of the interference wave, and the fifth OMP removes the influence of the interference wave almost completely. However, it takes more OMPs to remove the interference wave than in this embodiment.

[0126] Note that, for the sake of ease of understanding, Figures 32, 33A, and 33B show the reduction effect of subtracting interference waves after the first to fifth OMPs of each cycle, and in an actual additional comparative example, subtraction of interference waves is only performed after the fifth OMP of each cycle.

[0127] [1] As described above, the signal processing unit 7 calculates an approximate value of the chirp slope K based on a two-dimensional map in preprocessing, and calculates the coefficient vector x using a dictionary matrix based on the approximate value in OMP processing. Furthermore, the signal processing unit 7 estimates the waveform of the interference wave based on the coefficient vector x and the basis vector in the interference waveform calculation processing, and subtracts the waveform of the interference wave from the waveform of the received wave in the subtraction processing. Then, if the noise floor of the distance FFT spectrum obtained by FFT from the received wave is larger than a specified value, the signal processing unit 7 performs preprocessing again. In this way, even if there are multiple different types of interference waves, they can be individually removed.

[0128] It should be noted that the present embodiment and the first and second embodiments have similar configurations and operations, and therefore similar effects can be obtained.

[0129] (Fourth embodiment) Next, a fourth embodiment will be described. This embodiment is different from the first to third embodiments in the calculation timing of the approximate value of the chirp slope K. Specifically, the processor 71 of the signal processing unit 7 calculates the chirp slope K during an off-duty period.

[0130] This will be described with reference to Fig. 34. In the first to third embodiments, on-duty periods and off-duty periods alternate. The radar device 1 transmits a transmission wave from the transmission antenna 2 multiple times during each on-duty period, receives a reception wave multiple times at the reception antenna 3, and mixes the reception signal RX with the local signal LO. The signal processing unit 7 calculates a distance FFT spectrum based on a beat signal obtained from the mixed signal during the on-duty period.

[0131] During each off-duty period, the radar device 1 does not operate either the transmitting antenna 2 or the receiving antenna 3, does not generate a local signal LO, and does not calculate a distance FFT spectrum in the signal processing unit 7.

[0132] In this case, there are cases where no externally generated interference wave LB occurs, as in the upper part of FIG. 34, and cases where interference wave LB occurs repeatedly from successive off-duty periods to on-duty periods, as in the middle part.

[0133] Therefore, in this embodiment, the radar device 1 generates a local signal LO even during the off-duty period. However, the transmitting antenna 2 is not operated. In other words, no transmission wave is sent out. Also, the receiving antenna 3 is operated so as to be able to receive the interference wave LB. If the receiving antenna 3 is an array antenna having multiple antenna elements, one of the multiple antenna elements may be operated and the remaining antenna elements may not be operated.

[0134] As a result, during the off-duty period, the signal processing unit 7 receives a beat signal equivalent to an IQ signal obtained by mixing the received signal RX, which includes the interference wave LB but does not include the reflected wave, with the local signal LO.

[0135] Then, in this off-duty period, the signal processing unit 7 calculates approximate values ​​of the chirp slope K and interference start time Tb of the interference wave LB based on this beat signal in the same manner as in the first embodiment, and calculates the waveform of the interference wave by the OMP based on the calculation result. Specifically, steps S1800 to S1807 in Fig. 21 are executed. The value of the chirp slope K selected in the last step S1804 in this process is considered to be very close to the chirp slope K of the interference wave LB in the next on-duty period. This is because there is a high possibility that the interference wave LB in this off-duty period and the interference wave LB in the corresponding on-duty period have the same generation source.

[0136] Alternatively, the signal processing unit 7 may calculate the approximate values ​​of the chirp slope K and the interference start time Tb of the interference wave LB based on the beat signal during the off-duty period, as in the second and third embodiments, and calculate the waveform of the interference wave using the OMP algorithm based on the calculation result. Specifically, the processing of Fig. 22 and Fig. 28 may be executed. The value of the chirp slope K selected in the last step S1804 during this processing is considered to be very close to the chirp slope K of the interference wave LB during the next on-duty period.

[0137] The value of chirp slope K obtained in this manner is used as an approximate value of chirp slope K in the on-duty period following this off-duty period. That is, the signal processing unit 7 performs processing corresponding to Fig. 21, Fig. 22, and Fig. 28 in the on-duty period. However, the approximate value of chirp slope K to be estimated is the chirp slope K calculated as described above in the immediately preceding off-duty period.

[0138] [1] As described above, the signal processing unit 7 calculates an approximate value of the chirp slope K of the interference wave during the on-duty period, based on the interference wave received from the receiving antenna 3, during the off-duty period preceding the on-duty period in which the transmission wave is transmitted from the transmitting antenna 2.

[0139] In this way, OMP can be executed by utilizing the free time of the off-duty period, and more time can be secured for processing. Also, since OMP can be started in the on-duty period with the value of chirp slope K being very narrow, the number of OMPs can be further reduced. Note that the same effects can be obtained from the same configurations and operations in this embodiment and the first to third embodiments.

[0140] (Other embodiments) The present invention is not limited to the above-mentioned embodiment, and can be modified as appropriate. The above-mentioned embodiments are not unrelated to each other, and can be combined as appropriate, except when the combination is clearly impossible. In the above-mentioned embodiments, the elements constituting the embodiment are not necessarily essential, except when it is specifically stated that they are essential or when it is clearly considered essential in principle. In the above-mentioned embodiments, when the numbers, values, amounts, ranges, etc. of the components of the embodiment are mentioned, they are not limited to the specific numbers, except when it is specifically stated that they are essential or when it is clearly limited to a specific number in principle. In particular, when multiple values ​​are exemplified for a certain amount, it is also possible to adopt a value between the multiple values, except when it is specifically stated otherwise or when it is clearly impossible in principle. In the above-mentioned embodiments, when the shape, positional relationship, etc. of the components are mentioned, they are not limited to the shape, positional relationship, etc., except when it is specifically stated or when it is limited to a specific shape, positional relationship, etc. in principle. In addition, the present invention also allows various modifications and modifications within an equivalent range to the above-mentioned embodiments.

[0141] The signal processing unit 7 and the method thereof described in the present disclosure may be realized by a dedicated computer provided by configuring a processor and a memory programmed to execute one or more functions embodied in a computer program. Alternatively, the signal processing unit 7 and the method thereof described in the present disclosure may be realized by a dedicated computer provided by configuring a processor with one or more dedicated hardware logic circuits. Alternatively, the signal processing unit 7 and the method thereof described in the present disclosure may be realized by one or more dedicated computers configured by a combination of a processor and a memory programmed to execute one or more functions and a processor configured with one or more hardware logic circuits. Also, the computer program may be stored in a computer-readable non-transitory tangible recording medium as instructions executed by a computer. [Explanation of symbols]

[0142] 1. Radar equipment 2 transmitting antennas 3 Receiving antenna 7 Signal Processing Section

Claims

1. A radar device (1), a transmission signal generating unit (51) for generating a transmission signal (TX) which is an electrical signal having a predetermined frequency change corresponding to a frequency modulation mode of a transmission wave, the transmission wave being a frequency-modulated radar wave, so as to transmit the transmission wave from a transmission antenna (2); The radar wave received by the receiving antenna (3) is mixed with a local signal (LO) to produce a predetermined cutoff frequency F ＬＰＳ a signal processing unit (7) that executes an interference suppression process (S1800 to S1809) for suppressing the influence of an interference wave contained in the received wave based on a signal processing result for a beat signal obtained by filtering the received wave; Equipped with The interference suppression process includes: A waveform estimation process (S1800 to S1808) for estimating the waveform of the interference wave; A subtraction process (S1809) of subtracting the estimated waveform of the interference wave from the waveform of the received wave, The waveform estimation process includes: a parameter estimation process (S1800 to S1807) for estimating parameters in an equation expressing a waveform of the interference wave using parameters including an interference start time Tb at which the interference wave starts to ride on the beat signal, a chirp slope K which is a slope of a time change in the frequency of the interference wave, an amplitude A of the interference wave, and a phase θ of the interference wave; an interference waveform calculation process (S1808) for calculating a waveform of the interference wave based on the parameters estimated by the parameter estimation process; Including, The parameter estimation process includes: an OMP process (S1802 to S1806) for selecting one of the basis vectors that maximizes an inner product with the received waveform vector y from the dictionary matrix D, in order to calculate the coefficient vector x in an arithmetic equation y=Dx, which is represented by the product of a dictionary matrix D of N rows and M columns that is a set of basis vectors corresponding to a combination of the interference start time Tb and the chirp slope K, and a coefficient vector x of M rows and 1 column that includes the amplitude A and the phase θ, using an orthogonal matching pursuit algorithm so that the coefficient vector x includes only one non-zero component and the rest are zero components; A radar device comprising: a pre-processing step (S1800, S1801, S110 to S170) for calculating an approximate value of the chirp slope K before the OMP processing, and limiting the range of values ​​of the chirp slope K used in the basis vectors of the dictionary matrix D so that the range falls within a predetermined range including the calculated approximate value.

2. In the pre-processing, the signal processing unit performs a threshold test on the beat signal to identify the interference start time Tb and the interference end time Te at which the interference wave ends being carried on the beat signal, and uses the identified interference start time Tb, interference end time Te, and cutoff frequency F ＬＰＳ 2. The radar device according to claim 1, wherein the approximate value of the chirp slope K is calculated based on:

3. the beat signal is a complex waveform, 3. The radar device according to claim 2, wherein, in the pre-processing, the signal processing unit specifies, as the interference start time Tb, a timing at which an absolute value of the beat signal changes from a state smaller than a predetermined threshold TH to a state larger than the threshold TH, and specifies, as the interference end time Te, a timing at which the absolute value of the beat signal changes from a state larger than the threshold TH to a state smaller than the threshold TH.

4. The estimated value of the chirp slope K is K=2F ＬＰＦ () The radar device according to claim 3 , wherein the calculated value is expressed by:

5. 2. The radar device according to claim 1, wherein, in the preprocessing, the signal processing unit calculates a map of a power distribution of the beat signal for each time and for each frequency by performing a short-time Fourier transform on the beat signal, and calculates the approximate value of the chirp slope K based on the calculated map.

6. 6. The radar device according to claim 5, wherein, in the pre-processing, the signal processing unit repeats a process of recording a time and frequency pair at which maximum power is achieved in a row of a minimum frequency in the map, and then recording a time and frequency pair at which maximum power is achieved in a time column next to the last recorded time in the map, and then calculates the approximate value of the chirp slope K based on the recorded time and frequency pairs.

7. 6. The radar device according to claim 5, wherein, in the pre-processing, the signal processing unit performs a process of recording a time and frequency pair at which maximum power is achieved in a row of a maximum frequency in the map, and then repeats a process of recording a time and frequency pair at which maximum power is achieved in a time column next to the last recorded time in the map, and then calculates the approximate value of the chirp slope K based on the recorded time and frequency pairs.

8. 6. The radar device according to claim 5, wherein, in the pre-processing, the signal processing unit performs a process of recording a pair of time and frequency at which maximum power is realized in a first time sequence of the map, and then repeats the process of recording a pair of time and frequency at which maximum power is realized in a time sequence next to the last recorded time in the map, and then calculates the approximate value of the chirp slope K based on the recorded pairs of time and frequency.

9. 6. The radar device according to claim 5, wherein the signal processing unit calculates the approximate value of the chirp slope K based on the map in the pre-processing, calculates the coefficient vector x using the dictionary matrix D based on the approximate value in the OMP processing, estimates the waveform of the interference wave based on the coefficient vector x and the basis vectors in the interference waveform calculation processing, subtracts the waveform of the interference wave from the waveform of the received wave in the subtraction processing, and then performs the pre-processing again if a noise floor of a spectrum obtained by Fourier transform of the received wave is greater than a specified value.

10. 2. The radar device according to claim 1, wherein the signal processing unit calculates the approximate value of the chirp slope K of the interference wave during the on-duty period based on an interference wave received from the receiving antenna during an off-duty period in which the transmitting antenna is not used, prior to an on-duty period in which the transmission wave is transmitted from the transmitting antenna.

11. When a transmission wave, which is a frequency-modulated radar wave, is transmitted from a transmission antenna (2), a local signal (LO) is mixed with a reception wave, which is a radar wave received by a reception antenna (3), and a predetermined cutoff frequency F ＬＰＳ A program used in a signal processing unit (7) of a radar device that obtains a beat signal by filtering the signal causing the signal processing unit to execute an interference suppression process (S1800 to S1809) for suppressing the influence of an interference wave included in the received wave based on a signal processing result for the beat signal; The interference suppression process includes: A waveform estimation process (S1800 to S1808) for estimating the waveform of the interference wave; A subtraction process (S1809) of subtracting the estimated waveform of the interference wave from the waveform of the received wave, The waveform estimation process includes: a parameter estimation process (S1800 to S1807) for estimating parameters in an equation expressing a waveform of the interference wave using parameters including an interference start time Tb at which the interference wave starts to ride on the beat signal, a chirp slope K which is a slope of a time change in the frequency of the interference wave, an amplitude A of the interference wave, and a phase θ of the interference wave; an interference waveform calculation process (S1808) for calculating a waveform of the interference wave based on the parameters estimated by the parameter estimation process; Including, The parameter estimation process includes: an OMP process (S1802 to S1806) for selecting one of the basis vectors that maximizes an inner product with the received waveform vector y from the dictionary matrix D, in order to calculate the coefficient vector x in an arithmetic equation y=Dx, which is represented by the product of a dictionary matrix D of N rows and M columns that is a set of basis vectors corresponding to a combination of the interference start time Tb and the chirp slope K, and a coefficient vector x of M rows and 1 column that includes the amplitude A and the phase θ, using an orthogonal matching pursuit algorithm so that the coefficient vector x includes only one non-zero component and the rest are zero components; and pre-processing (S1800, S1801, S110 to S170) for calculating an approximate value of the chirp slope K before the OMP processing and limiting the range of values ​​of the chirp slope K used in the basis vectors of the dictionary matrix D so that the range falls within a predetermined range including the calculated approximate value.

Images