A nonlinear estimation and correction method for frequency modulated continuous wave radar
Through closed-loop testing and nonlinear signal processing technology, the nonlinear phase of RF frequency in frequency modulated continuous wave radar is estimated and corrected, and the phase error problem caused by nonlinear distortion is solved, and the distance measurement accuracy and imaging quality of the radar are improved, so as to achieve no additional need Efficient nonlinear correction effect of hardware.
Patent Information
- Application Number
- CN202210074842.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-01-21
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2042-01-21
AI Technical Summary
The nonlinear distortion in the linear frequency modulation signal of FM continuous wave radar introduces empty-displacement phase error, reducing signal quality, affecting the ranging accuracy and imaging quality.
By selecting a delay line of a predetermined length for closed-loop testing, the delay line difference frequency signal is obtained, the ideal difference frequency signal is constructed and mixed with it, the phase of the nonlinear signal is extracted, and the radio frequency nonlinear phase is estimated using upsampling and z-transform domain methods are used to perform Downsampling and smoothing are performed, and nonlinear correction is finally performed using matching Fourier transform.
It effectively suppresses the problems caused by nonlinearity in the signal, improves the ranging accuracy and imaging quality, enhances the detection ability of adjacent targets within the radar dynamic range, and eliminates the need for additional hardware equipment, reducing system complexity.
Smart Images

Figure CN114397650B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of signal processing, and in particular to a nonlinear estimation and correction method for a frequency modulated continuous wave radar. Background Art
[0002] Compared with traditional pulse radar, Frequency Modulated Continues Waveform (FMCW) radar has the advantages of low cost, higher distance resolution, stronger penetration ability, and the ability to eliminate distance blind spots. Therefore, FMCW radar is widely used in the field of high-precision ranging and imaging.
[0003] The most critical factor affecting the ranging and imaging accuracy of FMCW radar is the linearity of the linear frequency modulation signal. The FMCW signal has a large bandwidth and the signal lasts for the entire transmission cycle, so it is very difficult to maintain the linearity of the signal frequency modulation slope. Due to the non-ideal characteristics of each component of the radar system, the distortion of the signal frequency modulation - frequency modulation nonlinearity will be introduced to a greater or lesser extent. This nonlinear component in the linear frequency modulation signal will introduce a space-varying phase error after dechirp processing, reducing the signal quality, which will have a certain impact on the signal spectrum, such as main lobe broadening, main lobe offset, asymmetric distortion, paired echoes, and excessive sidelobe energy, which seriously affect the ranging accuracy and imaging quality. Summary of the invention
[0004] The main purpose of the present invention is to provide a nonlinear estimation and correction method for a frequency modulated continuous wave radar, so as to improve the ranging accuracy or imaging quality of the frequency modulated continuous wave radar.
[0005] To achieve the above object, the technical solution adopted by the present invention is: a nonlinear estimation and correction method for frequency modulated continuous wave radar, comprising:
[0006] Step 1, selecting a delay line of a predetermined length, performing a closed-loop test on the radar system, and obtaining a delay line difference frequency signal;
[0007] Step 2, obtaining the delay line difference frequency through the spectrum of the delay line difference frequency signal, and constructing an ideal difference frequency signal based on it, mixing the ideal difference frequency signal with the delay line difference frequency signal, and obtaining a nonlinear signal with RF nonlinearity and echo nonlinearity superimposed;
[0008] Step 3, extracting the phase of the nonlinear signal and unwrapping it to obtain the nonlinear phase in the delay line difference frequency signal;
[0009] Step 4: upsampling the nonlinear phase data according to a preset threshold, constructing a differential equation using the upsampled nonlinear phase, and solving the differential equation using a z-transform domain method to obtain an initial estimate of the RF nonlinear phase;
[0010] Step 5: downsampling the initial estimate of the RF nonlinear phase to restore the original data length, and filtering out the "burr" signal caused by the mismatch between the discrete delay and the continuous delay to obtain an estimate of the RF nonlinear phase;
[0011] Step 6: Apply the radio frequency nonlinear phase estimation to the actual detection data and perform nonlinear correction using matched Fourier transform to obtain detection data after nonlinear correction.
[0012] Furthermore, in the process of constructing the ideal difference frequency signal, the difference frequency of the delay line is set to f beat , then the ideal difference frequency signal constructed is:
[0013] s b,ideal (t) = exp(j2πf beat )
[0014] Furthermore, in the process of mixing the ideal difference frequency signal with the delay line difference frequency signal, the delay line difference frequency signal is assumed to be s b (t), the RF nonlinear phase is ε(t), and the delay line delay is τ, then the nonlinear signal is:
[0015] s e (t) = s b (t)·s b,ideal (t) = exp{j[ε(t)-ε(t-τ)]}
[0016] Furthermore, before obtaining the estimation of the radio frequency nonlinear phase, the method further includes:
[0017] According to a preset threshold, that is, a deviation between discrete delay and continuous delay is less than Δms, upsampling the nonlinear phase data;
[0018] Constructing a differential equation using the upsampled nonlinear phase, and solving the differential equation using a z-transform domain method to obtain an initial estimate of the radio frequency nonlinear phase;
[0019] Downsampling the initial estimate of the radio frequency nonlinear phase to restore the original data length;
[0020] Filter out "glitch" caused by the mismatch between discrete delay and continuous delay.
[0021] Furthermore, the nonlinear phase after upsampling is used to construct a differential equation, and the differential equation is solved using a z-transform domain method, and the nonlinear phase after upsampling is φ Δε [n], the constructed difference equation is:
[0022] φ Δε [n]=ε[n]-ε[nN τ ]
[0023] Where n is a discrete time variable, N τ is the discrete reference delay. Let the system response be Then the z-domain representation of the initial estimate of the RF nonlinear phase can be obtained:
[0024] ε first (z)=H(z)·φ Δε (z)
[0025] The ε first (z) performing an inverse z-transform to obtain an initial estimate of the RF nonlinear phase.
[0026] Furthermore, filtering out the “burrs” caused by the mismatch between discrete delay and continuous delay includes:
[0027] The initial estimate of the RF nonlinear phase after the downsampling is subjected to data smoothing. The length of the smoothing window cannot be too large. The length of the smoothing window should be set at about 2 to 5 "burr" cycle lengths to avoid smoothing out the periodic nonlinear components of the phase itself, and obtain an RF nonlinear phase estimate with the same length as the original data.
[0028] Furthermore, the preset threshold is 0.5 ms.
[0029] Furthermore, the length of the smoothing window is 3 sampling points.
[0030] Compared with the prior art, the present invention has the following beneficial effects:
[0031] (1) The present invention estimates and corrects the nonlinear phase of the frequency modulated continuous wave radar system based on a software method, does not require additional hardware equipment, can save hardware costs, reduce system complexity, and has wide applicability.
[0032] (2) The present invention uses a difference frequency equation filtering algorithm to estimate the nonlinear phase of a frequency modulated continuous wave radar system, and can estimate both periodic and non-periodic nonlinear phases simultaneously, is not limited by a single nonlinear model, and has a high estimation accuracy.
[0033] (3) The present invention uses a matching Fourier algorithm to perform nonlinear phase correction on the actual detection data of the frequency modulated continuous wave radar, and still has a good correction effect when the system nonlinearity is high.
[0034] (4) After correction, the present invention effectively suppresses the problems caused by nonlinearity in the signal, reduces the sidelobe energy, and enhances the detection capability of adjacent targets within the dynamic range of the radar. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.
[0036] Figure 1 A schematic flow chart of a nonlinear estimation and correction method for a frequency modulated continuous wave radar provided in one embodiment of the present invention;
[0037] Figure 2 The actual nonlinear phase of the radio frequency signal in the nonlinear estimation and correction method of the frequency modulated continuous wave radar provided in one embodiment of the present invention;
[0038] Figure 3 A normalized spectrum of a difference frequency signal in a nonlinear estimation and correction method for a frequency modulated continuous wave radar provided in an embodiment of the present invention;
[0039] Figure 4 The nonlinear estimation result and the "burr" phenomenon and the smoothing result in the nonlinear estimation and correction method of the FMCW radar provided in one embodiment of the present invention;
[0040] Figure 5 A comparison between a differential filtering method, a coherent integration method and actual nonlinearity in a nonlinearity estimation and correction method for a frequency modulated continuous wave radar provided in an embodiment of the present invention;
[0041] Figure 6 Comparison of estimation errors of the differential filtering method and the coherent integration method in the nonlinear estimation and correction method of the frequency modulated continuous wave radar provided in one embodiment of the present invention;
[0042] Figure 7 The normalized spectrum of the closed-loop data delay line difference frequency signal collected in the nonlinear estimation and correction method of the frequency modulated continuous wave radar provided in one embodiment of the present invention;
[0043] Figure 8The radio frequency nonlinear phase estimated from the collected closed-loop data in the nonlinear estimation and correction method of the frequency modulated continuous wave radar provided in one embodiment of the present invention;
[0044] Fig. 9 A normalized spectrum of another set of closed-loop data difference frequency signals after RF nonlinear phase correction using closed-loop data estimation in a nonlinear estimation and correction method for a frequency modulated continuous wave radar provided in an embodiment of the present invention;
[0045] Fig.10 The imaging result after correction of actual field detection data in the nonlinear estimation and correction method of the frequency modulated continuous wave radar provided in one embodiment of the present invention;
[0046] Fig.11 It is one of the data of the imaging results after correction of actual field detection data in the nonlinear estimation and correction method of FMCW radar provided in one embodiment of the present invention. DETAILED DESCRIPTION
[0047] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0048] See also Figure 1 , Figure 1 A flow chart of a nonlinear estimation and correction method for a frequency modulated continuous wave radar provided by an embodiment of the present invention, the method mainly comprises the following steps:
[0049] S101, selecting a delay line of appropriate length, performing a closed-loop test on the radar system, and obtaining a delay line difference frequency signal;
[0050] Use delay lines of different lengths to perform closed-loop testing, select relatively stable closed-loop data as the test data, and discard data with large deviations. Figure 2 As shown, the normalized spectrum of the collected closed-loop data difference frequency signal is as follows Figure 3 shown.
[0051] S102, obtaining the delay line difference frequency through the delay line difference frequency signal spectrum, and constructing an ideal difference frequency signal based on it, mixing the ideal difference frequency signal with the delay line difference frequency signal, and obtaining a nonlinear signal with radio frequency nonlinearity and echo nonlinearity superimposed;
[0052] Let the delay line difference frequency be f beat, then the ideal difference frequency signal constructed is:
[0053] s b,ideal (t) = exp(j2πf beat )
[0054] The delay line difference frequency signal is s b (t), the RF nonlinear phase is ε(t), and the delay line delay is τ, then the nonlinear signal is:
[0055] s e (t) = s b (t)·s b,ideal (t) = exp{j[ε(t)-ε(t-τ)]}
[0056] S103, extracting the phase of the nonlinear signal and unwrapping it to obtain the nonlinear phase in the difference frequency signal;
[0057] S104, upsampling the nonlinear phase data according to a preset threshold, constructing a differential equation using the upsampled nonlinear phase, and solving the differential equation using a z-transform domain method to obtain an initial estimate of the radio frequency nonlinear phase;
[0058] The threshold is set to 0.5ms, that is, the deviation between the discrete delay and the continuous delay does not exceed 0.5ms, the nonlinear phase is upsampled, and the differential equation is constructed: Δε [n]=ε[n]-ε[nN τ ], where n is a discrete time variable, N τ is the discrete reference delay. Let the system response be Then the z-domain representation of the initial estimate of the RF nonlinear phase can be obtained:
[0059] ε first (z)=H(z)·φ Δε (z)
[0060] The obtained ε first (z) performing an inverse z-transform to obtain an initial estimate of the RF nonlinear phase.
[0061] S105, downsampling the initial estimate of the radio frequency nonlinear phase to restore the original data length, and filtering out "burrs" caused by the mismatch between discrete delay and continuous delay to obtain an estimate of the radio frequency nonlinear phase;
[0062] The initial estimation result of the RF nonlinear phase after downsampling is as follows: Figure 4As shown in , data smoothing is performed. The smoothing window length cannot be too large. The length of the smoothing window should be set to about 2 to 5 "burr" cycle lengths to avoid smoothing out the periodic nonlinear components of the phase itself. The smoothing window length is set to 3 sampling points to obtain the same RF nonlinear phase estimation as the original data length. The final estimation result is shown in Figure 4 The method of the present invention is compared with the coherent integration method, and the comparison results are shown in FIG. Figure 5 As shown in the figure, the estimation error comparison results are as follows Figure 6 shown.
[0063] S106: Apply the nonlinear phase estimation to actual detection data and perform nonlinear correction using matched Fourier transform to obtain detection data after nonlinear correction.
[0064] The following is a specific example of the embodiment of the present invention:
[0065] First, a delay line with a length of about 6.45m is used for closed-loop testing to obtain a set of delay line difference frequency signal data, and then a delay line with a length of about 5.70m is used for closed-loop testing to obtain another set of delay line difference frequency signal data. The normalized spectrum of the 6.45m delay line difference frequency signal is shown in Figure 7 shown.
[0066] Then, the ideal difference frequency is obtained by using the delay line difference frequency signal data of 6.45m, and the ideal difference frequency signal is constructed and mixed with the delay line difference frequency signal to obtain a nonlinear signal.
[0067] Then, the obtained nonlinear signal phase is extracted and unwrapped to obtain the nonlinear phase in the delay line difference frequency signal.
[0068] Then, the nonlinear phase in the difference frequency signal is upsampled according to a preset threshold of 0.5 ms, so that the deviation between the discrete delay and the continuous delay is less than 0.5 ms.
[0069] Then, a differential equation is constructed according to the upsampled nonlinear phase, and a preliminary estimate of the RF nonlinear phase is obtained by solving it using the z-transform domain method. The preliminary estimate result is then downsampled to restore the original data length.
[0070] Due to the mismatch between discrete delay and continuous delay, the RF nonlinear phase estimation result will have "burrs". It is necessary to smooth the data of the preliminary estimation result. The smoothing window length is set to 3 sampling points. After smoothing, the RF nonlinear phase estimation result is obtained, such as Figure 8 shown.
[0071] Then, the nonlinear phase estimated from the closed-loop data of the 6.45m delay line is used to perform nonlinear correction on the closed-loop data of another delay line. Fig. 9 shown.
[0072] Then, the RF nonlinear phase estimated by the closed-loop test is used to match the actual field detection data with the Fourier transform nonlinear correction. The image after nonlinear correction is as follows: Fig.10 Extract one of the data, and the comparison results before and after correction are shown as follows Fig.11 As shown, the sidelobe energy is suppressed, which effectively suppresses the problems caused by nonlinearity in the signal and enhances the detection capability of adjacent targets within the dynamic range of the radar.
[0073] In the example of the present invention, a delay line of suitable length is selected, a closed-loop test is performed on the radar system, a delay line difference frequency signal is obtained, the delay line difference frequency frequency is obtained through the difference frequency signal spectrum, and an ideal difference frequency signal is constructed by this, the ideal difference frequency signal is mixed with the delay line difference frequency signal, a nonlinear signal of RF nonlinearity and echo nonlinearity superposition is obtained, the phase of the nonlinear signal is extracted and unwrapped, and the nonlinear phase in the difference frequency signal is obtained, the nonlinear phase data is upsampled according to a preset threshold, a differential equation is constructed using the upsampled nonlinear phase, and the differential equation is solved using the z-transform domain method to obtain an initial estimate of the RF nonlinear phase, the initial estimate of the RF nonlinear phase is downsampled to restore the original data length, and the "burrs" caused by the mismatch between discrete delay and continuous delay are filtered out to obtain an estimate of the RF nonlinear phase, the nonlinear phase is applied to actual detection data and nonlinear correction is performed using a matched Fourier transform to obtain detection data after nonlinear correction. After correction, the problems caused by nonlinearity in the signal are effectively suppressed, the sidelobe energy is reduced, and the detection capability of adjacent targets within the dynamic range of the radar is enhanced.
[0074] It should be noted that, for the convenience of description, the aforementioned method embodiments are all described as a series of action combinations, but those skilled in the art should be aware that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.
[0075] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0076] The above is a description of the nonlinear estimation and correction method of the FMCW radar provided by the present invention. For those skilled in the art, according to the concept of the embodiments of the present invention, there may be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present invention.
Claims
1. A nonlinear estimation and correction method for frequency modulated continuous wave radar, characterized in that: The steps include: Step 1, selecting a delay line of a predetermined length, performing a closed-loop test on the radar system, and obtaining a delay line difference frequency signal; Step 2, obtaining the delay line difference frequency through the spectrum of the delay line difference frequency signal, and constructing an ideal difference frequency signal based on it, mixing the ideal difference frequency signal with the delay line difference frequency signal, and obtaining a nonlinear signal with RF nonlinearity and echo nonlinearity superimposed; Step 3, extracting the phase of the nonlinear signal and unwrapping it to obtain the nonlinear phase in the delay line difference frequency signal; Step 4: upsampling the nonlinear phase data according to a preset threshold, constructing a differential equation using the upsampled nonlinear phase, and solving the differential equation using a z-transform domain method to obtain an initial estimate of the RF nonlinear phase; Step 5: downsampling the initial estimate of the RF nonlinear phase to restore the original data length, and filtering out the "burr" signal caused by the mismatch between the discrete delay and the continuous delay to obtain an estimate of the RF nonlinear phase; Step 6: Apply the radio frequency nonlinear phase estimation to the actual detection data and perform nonlinear correction using matched Fourier transform to obtain detection data after nonlinear correction; In step 2, in the process of constructing the ideal difference frequency signal, the delay line difference frequency is set to , then the ideal difference frequency signal constructed is: , Among them, exp is the exponential function; In step 2, during the mixing process of the ideal difference frequency signal and the delay line difference frequency signal, the delay line difference frequency signal is set to , the RF nonlinear phase is , the delay line delay is , then the nonlinear signal is: 。 2. The nonlinear estimation and correction method of FMCW radar according to claim 1, characterized in that: In the step 4, before obtaining the estimation of the radio frequency nonlinear phase, the method further includes: According to the preset threshold, the deviation between discrete delay and continuous delay is less than ms, upsampling the nonlinear phase data; Constructing a differential equation using the upsampled nonlinear phase, and solving the differential equation using a z-transform domain method to obtain an initial estimate of the radio frequency nonlinear phase; Downsampling the initial estimate of the radio frequency nonlinear phase to restore the original data length; Filter out "glitch" caused by the mismatch between discrete delay and continuous delay.
3. The nonlinear estimation and correction method of FMCW radar according to claim 2, characterized in that: The nonlinear phase after upsampling is used to construct a differential equation, and the differential equation is solved using the z-transform domain method. The nonlinear phase after upsampling is set to , then the constructed difference equation is: , in, is a discrete time variable, is the discrete reference delay, and the system response is , the z-domain representation of the initial estimate of the RF nonlinear phase can be obtained: , The An inverse z-transform is performed to obtain an initial estimate of the radio frequency nonlinear phase.
4. The nonlinear estimation and correction method of FMCW radar according to claim 2, characterized in that: In step 5, filtering out "burrs" caused by the mismatch between discrete delay and continuous delay includes: The initial estimate of the RF nonlinear phase after downsampling is subjected to data smoothing, and the length of the smoothing window should be set at 2 to 5 "burr" cycle lengths to avoid smoothing out the periodic nonlinear components of the phase itself, and obtain an RF nonlinear phase estimate with the same length as the original data.
5. The nonlinear estimation and correction method of FMCW radar according to claim 2, characterized in that: The preset threshold is 0.5 ms.
6. The nonlinear estimation and correction method of FMCW radar according to claim 4, characterized in that: The length of the smoothing window is 3 sampling points.
