A method for measuring the distance of a level radar based on orthogonal matching pursuit

Abstract

A method for measuring the distance of a level radar based on orthogonal matching pursuit belongs to the field of radar level distance measurement. It solves the problems of poor measurement accuracy when using existing radar level meters to measure distance and high computational complexity in existing methods for improving the accuracy of measuring distance. The present invention uses a level radar to transmit a linear frequency modulation signal in the direction to be measured and receive the echo signal; obtains the difference frequency signal; intercepts and samples the difference frequency signal to form a sampling sequence of Q×N elements and converts it into an underdetermined matrix; uses the information theory method to estimate the number of targets K of the underdetermined matrix based on the likelihood function; uses the OMP algorithm to solve the underdetermined matrix to obtain the difference frequencies of K targets; calculates the relative distance information between the K targets and the radar. The present invention is applicable to radar level distance measurement.

Description

Technical Field

[0001] The present invention belongs to the field of radar level distance measurement. Background Art

[0002] A radar level gauge is a non-contact device for measuring the relative distance of an object (solid or liquid) from the radar. Currently, it is widely used in industrial fields such as high-temperature, high-pressure, corrosive solid / liquid measurement, oil tank liquid level, food processing, etc. The radar level gauge is based on the mechanism of electromagnetic wave propagation. The radar emits electromagnetic waves. Since the free space and the object surface have different dielectric constants, part of the electromagnetic wave energy will be reflected back and received by the radar after hitting the object surface. There is a time delay between the received echo and the transmitted echo signal, and this time delay characterizes the distance information of the object relative to the radar, that is, the level information.

[0003] Currently, the commonly used transmitted waveform of the radar level gauge is the linear frequency modulation continuous wave. By using the characteristics of large bandwidth, low power consumption, and safety of the linear frequency modulation continuous wave, the non-contact ranging problem can be effectively solved. The ranging principle is that the radar transmitted signal is mixed with the echo signal received after a certain time delay to obtain a difference frequency signal, and then it is transformed to the frequency domain by using the discrete Fourier transform. The frequency value corresponding to the spectral peak at this time is found through spectral peak search, which is the difference frequency. This frequency can obtain the level information through the distance-frequency conversion formula. With the continuous development of industry, more and more application scenarios have requirements for ranging accuracy. Because under the condition of a certain bandwidth, the distance resolution of the target is fixed. How to break through this distance limitation under effective hardware conditions is the key problem to be solved.

[0004] Currently, the commonly used methods to improve the distance measurement accuracy include: (1) Increasing the AD data sampling rate can increase the amount of time-domain signal information obtained. After discrete Fourier transform, the frequency-domain spectral resolution is improved, and at the same time, the ranging accuracy and distance resolution are also improved. However, the disadvantages are also very obvious, which will lead to an increase in the number of calculation points, that is, it is necessary to increase the storage space of the hardware and also increase the calculation complexity. (2) Using the CZT algorithm, CZT is a spectral result estimated based on the discrete Fourier transform. It is a method to improve the distance estimation accuracy by fitting the position of the spectral peak without increasing the number of FFT calculation points by using the data fitting method. This method can improve the distance estimation accuracy, but cannot improve the distance resolution, and this method is based on FFT, and 2 additional FFT calculations and 1 inverse FFT calculation are required according to the number of fitting points. Therefore, it will inevitably increase the calculation complexity. Summary of the Invention

[0005] The purpose of the present invention is to solve the problems of poor measurement accuracy when using a radar level gauge to measure distance and high calculation complexity in the existing methods for improving the measurement distance accuracy, and a level radar ranging method based on orthogonal matching pursuit is proposed.

[0006] The present invention provides a method for measuring distance of a position radar based on orthogonal matching pursuit, comprising:

[0007] Step 1: Use the level radar to transmit a linear frequency modulation signal to the direction to be measured and receive the echo signal;

[0008] Step 2: conjugate and multiply the linear frequency modulation signal emitted by the level radar with the echo signal to obtain a difference frequency signal;

[0009] Step 3: intercept Q frequency sweeping cycles of the difference frequency signal, sample each frequency sweeping cycle N times, form a sampling sequence of Q×N elements, and convert the sampling sequence into a Q×N dimensional underdetermined matrix; wherein, are all positive integers, and Q<N;

[0010] Step 4: Use information theory to estimate the target number K of the underdetermined matrix based on the likelihood function;

[0011] Step 5: According to the number of targets K in the underdetermined matrix, the OMP algorithm is used to solve the underdetermined matrix to obtain the difference frequencies of the K targets;

[0012] Step 6: Calculate the relative distance information between the K targets and the radar based on the difference frequencies of the K targets obtained in step 5.

[0013] Further, in the present invention, in step 1, the linear frequency modulation signal is S T (t):

[0014]

[0015] Where B / T is the linear frequency modulation signal S T (t), B is the frequency modulation slope, T is the frequency sweep period, t is the time, j is the imaginary unit, and f0 is the starting frequency of the frequency sweep.

[0016] Furthermore, in the present invention, in step 1, the received echo signal is S R (t):

[0017]

[0018] Among them, τ = 2R / c, which represents the echo delay of the signal, R represents the distance of the object relative to the radar, and c = 3×10 8 m / s represents the propagation speed of electromagnetic waves in free space.

[0019] Furthermore, in the present invention, in step 2, the difference frequency signal obtained is x(t):

[0020]

[0021] Where: x(t) is the difference frequency signal, The linear frequency modulation signal S transmitted by the radar T (t) conjugate.

[0022] Furthermore, in the present invention, in step three, the nth sampling signal among the N sampling signals is x(n):

[0023]

[0024] S T * (n) represents the conjugate of the linear frequency modulation signal of the nth sampling, S R (n) is the echo signal of the nth sampling, simplified as:

[0025]

[0026] Among them, N = f s T, f s is the sampling rate, n b (n) is the noise signal at the nth sampling point, f b is the difference frequency, S b = exp(jφ b ), φ b is the phase of the difference frequency signal, φ b = -2πf0t.

[0027] Furthermore, in the present invention, in step three, the sampling signal is converted into an underdetermined matrix as:

[0028] X = AS + N0

[0029] Among them, X is the received signal matrix constructed by the difference frequency signal; A is the constructed overcomplete redundant atom dictionary, A is a matrix composed of Q sweep frequency periods, each period with N sampling point data, expressed as: f i represents the ith sampling frequency, the superscript symbol T represents the transpose of the matrix, represents the echo signal amplitude corresponding to N sampling points, S is a sparse or compressible signal, it only has larger amplitudes at a limited number of frequency points, and according to the sparse representation theory, the corresponding limited number of frequencies are the frequency values we want to estimate; represents additive white Gaussian noise, represents a complex signal, and the dimensions are Q×N dimension and Q×1 dimension respectively.

[0030] Furthermore, in the present invention, in step five, the specific method for obtaining the difference frequencies of K targets is:

[0031] First: Set the input of the OMP algorithm: matrix X, overcomplete redundant atom dictionary A, and target number K;

[0032] Initialize the OMP algorithm: residual r0 = X, index set Λ0 = Φ (indicating an empty set), and iteration initial value k = 1;

[0033] Then execute:

[0034] Step A1: Find the atom number in dictionary A that has the highest similarity with the amplitude S of the echo signal: Among them, r k-1 represents the changing residual in the k-th iteration, and a i represents the i-th column in dictionary A;

[0035] Step A2: Add the atom number with the highest similarity found to the index set Λ k : Λ k = Λ k-1 ∪{λ k}Λ k 、Λ k-1 respectively represent the index sets after the k-th iteration, and λ k represents the atom number calculated in step A1 during the k-th iteration.

[0036] Step A3: Update the residual using the index set Λ k : Among them, represents the matrix composed of the column vectors of dictionary A corresponding to the index set Λ k during the k-th iteration;

[0037] Step A4: Let k = k + 1. If k < K, return to step A1 until k = K, and the index set Λ k = Λ k-1 ;

[0038] Step A5: Take the frequency k corresponding to the index set Λ as the difference frequency f b (k) of the k-th target, where k = 1,..., K.

[0039] Furthermore, in the present invention, in step six, among the relative distance information between the K targets and the radar, the relative distance between the k-th target and the radar is:

[0040]

[0041] Among them, R dis (k) represents the distance of the k-th target, and f b (k) represents the difference frequency of the k-th target searched in the OMP algorithm.

[0042] The present invention applies the OMP technology to the radar ranging mathematical model, effectively improving the distance measurement accuracy of the radar level gauge, avoiding the fixed distance resolution problem caused by FFT calculation, constructing an overcomplete redundant dictionary using the difference frequency signal, applying the orthogonal matching pursuit technology to the radar distance resolution mathematical model, and solving it with the OMP algorithm based on the optimization of underdetermined equations. The position information of the target can be obtained through a limited number of iterative calculations, effectively reducing the amount of calculation, and there is also a certain improvement in accuracy compared with the FFT algorithm. Therefore, the distance measurement accuracy of the radar level gauge is effectively improved. Brief Description of the Drawings

[0043] Figure 1 is the flowchart of the method of the present invention;

[0044] Figure 2 is the comparison chart of the level distance obtained by using the method of the present invention and the level distance obtained by using the existing FFT algorithm. Detailed Embodiments

[0045] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0046] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0047] Detailed Embodiment 1: Next, in conjunction with Figure 1 this embodiment will be described. The method for measuring the level of a radar based on the orthogonal matching pursuit technology, a method for measuring the level of a radar based on the orthogonal matching pursuit, includes:

[0048] Step 1: Use the level radar to transmit a linear frequency modulated signal in the direction to be measured and receive the echo signal;

[0049] The linear frequency modulated signal is S T (t):

[0050]

[0051] where B / T is the frequency modulation slope of the linear frequency modulated signal S T (t), B is the bandwidth, T is the sweep period, t is the time, j is the imaginary unit, and f0 is the starting frequency of the frequency sweep;

[0052] The received echo signal is S R (t):

[0053]

[0054] Among them, τ = 2R / c represents the echo time delay of the signal, R represents the distance of the object relative to the radar, and c = 3×10 8 m / s represents the propagation speed of electromagnetic waves in free space;

[0055] Step 2: Conjugate multiply the linear frequency modulation signal emitted by the level radar and the echo signal to obtain the difference frequency signal;

[0056] The obtained difference frequency signal is x(t):

[0057]

[0058] Where: x(t) is the difference frequency signal, is the conjugate of the linear frequency modulation signal S T (t);

[0059] Step 3: Intercept Q sweep cycles of the difference frequency signal, sample N times for each sweep cycle, form a sampling sequence of Q×N elements, and convert the sampling sequence into a Q×N dimensional underdetermined matrix; where, are all positive integers, and Q < N;

[0060] The nth sampling signal in the sampling sequence is x(n):

[0061]

[0062] S T * (n) represents the conjugate of the linear frequency modulation signal at the nth sampling, and S R (n) is the echo signal at the nth sampling, which is simplified to:

[0063]

[0064] N = f s T, where f s is the sampling rate, n b (n) is the noise signal at the nth sampling point, f b is the difference frequency, φ b is the phase of the difference frequency signal, and φ b = -2πf0t;

[0065] The Q×N dimensional underdetermined matrix converted from the sampling sequence is:

[0066] X = AS + N0

[0067] Where, X is the received signal matrix constructed from the difference frequency signal; A is the constructed over-complete redundant atom dictionary, and A is a matrix composed of data of Q sweep frequency periods, with N sampling points in each period, expressed as: f i represents the i-th sampling frequency, and the superscript symbol T represents the transpose of the matrix. represents the echo signal amplitude corresponding to N sampling points. S is a sparse or compressible signal, which only has relatively large amplitudes at a limited number of frequency points. According to the sparse representation theory, the corresponding limited number of frequencies are the frequency values we need to estimate. represents additive white Gaussian noise. represents a complex signal, and the dimensions are Q×N and Q×1 respectively.

[0068] Step Four: Use the information theory method to estimate the target number K of the underdetermined matrix based on the likelihood function.

[0069] Step Five: According to the target number K of the underdetermined matrix, use the OMP algorithm to solve the underdetermined matrix and obtain the difference frequencies of K targets.

[0070] The specific method for obtaining the difference frequencies of K targets is as follows:

[0071] First: Set the input of the OMP algorithm: matrix X, over-complete redundant atom dictionary A, and target number K.

[0072] Initialize the OMP algorithm: residual r0 = X, index set Λ0 = Φ (indicating an empty set), and iteration initial value k = 1.

[0073] Then execute:

[0074] Step A1: Find the atom number in dictionary A that has the highest similarity with the echo signal amplitude S: where r k-1 represents the changing residual in the k-th iteration, and a i represents the i-th column in dictionary A.

[0075] Step A2: Add the found atom number with the highest similarity to the index set Λ k : Λ k = Λ k-1 ∪{λ k} Λ k 、Λ k-1 respectively represent the index sets after the k-th iteration, and λ k represents the atom number calculated in Step A1 during the k-th iteration.

[0076] Step A3: Update the residual using the index set Λ k : where, Denote the index set Λ at the k-th iteration k The matrix composed of the column vectors of the corresponding dictionary A;

[0077] Step A4. Let k = k + 1. If k < K, return to Step A1 until k = K, and the index set Λ k = Λ k-1 ;

[0078] Step A5. Take the frequency k corresponding to the index set Λ as the difference frequency f b (k) of the k-th target, where k = 1, …, K;

[0079] Step Six. Calculate the relative distance information between the K targets and the radar according to the difference frequencies of the K targets obtained in Step Five;

[0080] Among the relative distance information between the K targets and the radar, the relative distance between the k-th target and the radar is:

[0081]

[0082] where R dis (k) represents the distance of the k-th target, and f b (k) represents the difference frequency of the k-th target searched in the OMP algorithm.

[0083] In this embodiment, a high-precision ranging method for a level radar based on the orthogonal matching pursuit technique is provided. Figure 1 It is a flowchart of distance estimation based on the OMP technique provided by the embodiment of the present invention, as Figure 1 shown, including:

[0084] (1) Obtain the radar echo zero intermediate frequency signal;

[0085] (2) Process the zero intermediate frequency signal to construct the received signal matrix X;

[0086] (3) Intercept M sweep cycles of the received signal, and construct an overcomplete redundant atom dictionary A with N sampling points in each cycle;

[0087] (4) Based on searching for the extreme value using the maximum likelihood theory, determine the number of targets K to be estimated;

[0088] (5) Use OMP to iteratively calculate the difference frequency index set Λ k , and obtain the difference frequencies f b (k) of the K targets;

[0089] (6) Calculate the target distance,

[0090] In this embodiment, taking two targets of 7.256 m and 8.991 m as examples, the results are obtained through the FFT and OMP algorithms as Figure 2 shown. The results show that the OMP algorithm can accurately estimate the true position of the target, while there will be a deviation in the estimation process of the FFT algorithm, and the distance estimation accuracy is worse than that of the OMP method.

[0091] Although the present invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed, as long as they do not depart from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the different dependent claims and the features described herein can be combined in a manner different from that described in the original claims. It should also be understood that the features described in connection with a single embodiment can be used in other described embodiments.

Claims

1. A method for measuring the level radar ranging based on orthogonal matching pursuit, characterized in that, Including: Step 1: Use a level radar to transmit a linear frequency modulation signal in the direction to be measured and receive the echo signal; Step 2: Conjugate multiply the linear frequency modulation signal transmitted by the level radar and the echo signal to obtain a difference frequency signal; Step 3: Intercept Q sweep cycles of the difference frequency signal, sample N times for each sweep cycle, form a sampling sequence of Q×N elements, and convert the sampling sequence into a Q×N dimensional underdetermined matrix; where both are positive integers and Q < N; Step 4: Use the information theory method to estimate the number of targets K of the underdetermined matrix based on the likelihood function; Step 5: According to the number of targets K of the underdetermined matrix, use the OMP algorithm to solve the underdetermined matrix to obtain the difference frequencies of K targets; Step 6: Calculate the relative distance information between K targets and the radar according to the difference frequencies of K targets obtained in Step 5.

2. A method for measuring the level radar ranging based on orthogonal matching pursuit according to claim 1, characterized in that In step one, the chirp signal is S T (t): where, B / T is the frequency modulation slope of the chirp signal S T (t), B is the bandwidth, T is the sweep period, t is the time, j is the imaginary unit, and f0 is the starting frequency of the frequency sweep.

3. The method for measuring the level radar ranging based on orthogonal matching pursuit according to claim 2, wherein In step one, the received echo signal is S R (t): Among them, τ = 2R / c represents the echo time delay of the signal, R represents the distance of the object relative to the radar, and c = 3×10 8 m / s represents the propagation speed of electromagnetic waves in free space.

4. A method for measuring the level radar ranging based on orthogonal matching pursuit according to claim 3, characterized in that In Step 2, the obtained difference frequency signal is x(t): where: x(t) is the difference frequency signal, is the conjugate of the chirp signal S T (t) transmitted by the radar.

5. A method for measuring the distance of a level radar based on orthogonal matching pursuit according to claim 4, characterized in that, In Step 3, the nth sampling signal of the sampling sequence is x(n): S T * (n) represents the conjugate of the chirp signal for the nth sampling, S R (n) is the echo signal for the nth sampling, which is simplified to: N = f s T, where f s is the sampling rate, n b (n) is the noise signal at the nth sampling point, f b is the difference frequency, φ b is the phase of the difference frequency signal, φ b = -2πf0t.

6. The method for measuring the level radar ranging based on orthogonal matching pursuit according to claim 5, characterized in that, In Step 3, the Q×N dimensional underdetermined matrix converted from the sampling sequence is: X = AS + N0 Among them, X is a received signal matrix constructed from the difference frequency signal; A is an overcomplete redundant atom dictionary constructed, and A is a matrix composed of data of Q sweep frequency periods, with N sampling points in each period, expressed as: f i represents the i-th sampling frequency, and the superscript symbol T represents the transpose of the matrix. represents the echo signal amplitude corresponding to N sampling points. represents additive white Gaussian noise. represents a complex signal, and the dimensions are Q×N dimension and Q×1 dimension respectively.

7. A method for measuring the level radar ranging based on orthogonal matching pursuit according to claim 6, characterized in that, In Step 5, the specific method for obtaining the difference frequencies of K targets is: First: Set the input of the OMP algorithm: matrix X, overcomplete redundant atom dictionary A, number of targets K; Initialize the OMP algorithm: residual r0 = X, index set Λ0 = Φ, iteration initial value k = 1; Then execute: Step A1: Find the atom number in dictionary A that has the highest similarity with the echo signal amplitude S: where r k-1 represents the residual that changes in the k-th iteration, and a i represents the i-th column in dictionary A; Step A2. Add the atom number with the maximum similarity found to the index set Λ k : Λ k = Λ k-1 ∪{λ k}, where Λ k and Λ k-1 represent the index sets after the k-th and (k - 1)-th iterations respectively, and λ k represents the atom number calculated in Step A1 during the k-th iteration; Step A3: Utilize the index set Λ k Update the residual: where represents the matrix composed of the column vectors of the dictionary A corresponding to the index set Λ at the k-th iteration; k ​ Step A4. Let k = k + 1. If k < K, return to Step A1 until k = K, and the index set Λ k = Λ k-1 ; Step A5: Take the index set Λ k corresponding frequencies as the difference frequency f b (k) of the k-th target, where k = 1, …, K.

8. A method for measuring the level radar ranging based on orthogonal matching pursuit according to claim 5, characterized in that, In Step 6, in the relative distance information between K targets and the radar, the relative distance between the kth target and the radar is: where R dis (k) represents the distance of the k-th target, and f b (k) represents the difference frequency of the k-th target searched in the OMP algorithm.

Images