Extended acceleration estimation range based on PRVAD

Abstract

This invention provides a method for extending the acceleration estimation range based on PRVAD, comprising: obtaining... K The intermediate frequency echo data of the target is mixed and compressed to obtain the time-domain echo signal. A fast-time Fourier transform is performed on the time-domain echo signal to obtain the fast-time frequency-domain signal. The autocorrelation of the fast-time frequency-domain signal is obtained. A variable-scale non-uniform Fourier transform is performed on the correlation delay time in the autocorrelation of the fast-time frequency-domain signal to achieve coherent accumulation of the correlation delay time. Demodulation processing is used to eliminate the influence of slow-time variables in the coherent accumulation of the correlation delay time. Blind speed compensation is performed on the signal after eliminating the influence of slow-time variables. Fourier transform decoupling is performed along the slow time to achieve energy accumulation of the correlation time. An inverse Fourier transform is performed on the fast-time frequency to achieve coherent accumulation with respect to the range frequency, obtaining the target range, velocity, and acceleration information. The estimated velocity and acceleration are used to compensate for the obtained velocity and acceleration.

Description

Technical Field

[0001] This invention relates to radar target tracking technology, and in particular to a method for extending the acceleration estimation range based on PRVAD. Background Technology

[0002] Long-term coherent accumulation (LTC) is one way to improve radar detection performance. Essentially, it increases the radar's observation time over the target area, thereby increasing the number of received echo pulses. The energy of multiple echo signals is then superimposed to improve the echo signal-to-noise ratio. LTC is the core of high-maneuverability target detection. However, during the observation period, the target's maneuverability can lead to cross-range cell migration and cross-Doppler cell migration problems in the target echo signal.

[0003] The methods to solve the above problems are: (1) The algorithm with the best detection performance for uniformly accelerated targets is the Generalized Radon-Fourier Transform (GRFT) algorithm, which is a high-mobility target detection algorithm close to the ideal state; however, the GRFT algorithm is implemented through three-dimensional parameter search, which has high computational complexity and is not conducive to engineering implementation. (2) The first-order Keystone Transform (KT) can effectively eliminate the distance movement caused by the target velocity, but the distance movement correction and acceleration estimation algorithm based on KT has a front-to-back relationship in the compensation form, which may cause the distance movement correction to affect the Doppler spread, and the KT transform has a large performance loss compared with the ideal compensation algorithm. (3) In 2019, Zhao Langxu proposed the PRVAD algorithm (Parameterized Range-Velocity-Acceleration Distribution) based on three-dimensional coherent accumulation. Under the combined effect of parameterization and three-dimensional coherent processing, the PRVAD algorithm has high accumulation performance. The original PRVAD algorithm may have the problem that the low radar pulse repetition frequency leads to a small acceleration estimation range, and the target acceleration is likely to exceed this range. Summary of the Invention

[0004] The purpose of this invention is to provide a method for extending the acceleration estimation range based on PRVAD, comprising:

[0005] Step S100: Obtain the intermediate frequency echo data of K targets, and perform down-conversion mixing filtering and pulse compression to obtain the time domain echo signal. Perform fast-time Fourier transform on the time domain echo signal to obtain the fast-time frequency domain signal.

[0006] Step S200: Let k=1, apply variable delay and constant delay to the fast time frequency domain signal in the slow time dimension, and obtain the autocorrelation of the fast time frequency domain signal;

[0007] Step S300: Perform a variable-scale non-uniform Fourier transform on the correlation delay time in the autocorrelation of the fast time-frequency domain signal to achieve coherent accumulation of the correlation delay time;

[0008] Step S400: De-line adjustment processing is used to eliminate the influence of slow time variables in the coherent accumulation of related delay times;

[0009] Step S500: Perform blind speed compensation on the signal that eliminates the influence of slow time variables;

[0010] Step S600: Perform Fourier transform on the signal after blind speed compensation along the slow time to decouple it and achieve energy accumulation for the relevant time.

[0011] Step S700: Apply inverse Fourier transform to the fast time frequency to achieve coherent accumulation of the distance frequency, and obtain the target distance, velocity, and acceleration information;

[0012] Step S800: Compensate the obtained velocity and acceleration with the estimated velocity and acceleration;

[0013] If k < K, proceed to step S200, k = k + 1, and replace the target distance, velocity, and acceleration in the fast time-frequency domain signal; if k = K, output the compensated acceleration.

[0014] Further, in step S200, the autocorrelation function of the fast time-frequency domain signal is obtained. As shown in formula (1)

[0015] (1)

[0016] in, For delayed variables, The delay time is constant. B represents the amplitude of the signal, and B represents the bandwidth. t represents wavelength. m Let c represent the speed of light, and k represent the k-th target. Let a represent the initial distance to the k-th target. k Let v represent the initial acceleration of the k-th target. k Let f represent the initial velocity of the k-th target. c C represents the carrier frequency. TDPSIAF () indicates a cross term.

[0017] Furthermore, in step S300, a variable-scale non-uniform Fourier transform is performed on the relevant delay time according to equation (2) to achieve coherent accumulation of the relevant delay time.

[0018] (2)

[0019] in, Indicates corresponding to The frequency variable, It is an introduced scaling factor. Indicates the amplitude of the signal after transformation. Indicates the intersection of terms. This is the Dirac function.

[0020] Furthermore, the influence of slow-time variables is eliminated according to equation (3).

[0021] (3)

[0022] in, Indicates the amplitude of the signal after transformation. Indicates the intersection of terms.

[0023] Furthermore, the blind speed compensation process in step S500 includes:

[0024] Will and Substituting into equation (3), we get

[0025] (4)

[0026] in, and Let X represent the autocorrelation radial baseband velocity and the autocorrelation radial velocity folding factor of the k-th target, respectively. f represents the autocorrelation radial blind velocity. p Indicates the pulse repetition period. The mismatch term representing the velocity folding factor;

[0027] The signal is searched using the velocity folding number method. Perform blind speed compensation, get

[0028] (5).

[0029] Furthermore, energy accumulation is achieved through a one-step variable-scale Fourier transform, as shown in equation (6).

[0030] (6)

[0031] in, Indicates corresponding to The frequency variable, The mismatch term representing the velocity folding factor, Indicates the intersection of terms.

[0032] Furthermore, in step S700, formula (7) is used to perform inverse Fourier transform on the fast time frequency f to achieve coherent accumulation with respect to the distance frequency f.

[0033] (7)

[0034] in, This represents the time variable corresponding to the distance frequency f. The mismatch term representing the velocity folding factor, Indicates the intersection of terms.

[0035] Furthermore, the velocity and acceleration are compensated according to equation (8).

[0036] (8)

[0037] in, Indicates the Doppler frequency. Indicates Fourier transform, This represents the inverse Fourier transform.

[0038] This invention improves and updates the original PRVAD algorithm by introducing a scaling factor in the delay time step and extending the delay time frequency, thereby expanding the acceleration measurement range of the PRVAD algorithm. This effectively solves the problem of missed detection of highly maneuverable targets when the target acceleration exceeds the measurement range of the radar system parameters.

[0039] The present invention will now be further described with reference to the accompanying drawings. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0041] Figure 2 This is a comparison chart of the acceleration estimation results before and after the improvement of the original PRVAD algorithm and the algorithm of this invention.

[0042] Figure 3 This is a schematic diagram of the coherent accumulation of the original PRVAD algorithm.

[0043] Figure 4 This is a schematic diagram of the coherent accumulation of the improved PRVAD algorithm. Detailed Implementation

[0044] An extended acceleration estimation range method based on PRVAD includes:

[0045] Step S100: Compress the time-domain echo signals of the pulses from the K targets. Fast-time Fourier transform is performed to obtain the fast-time frequency domain signal. ;

[0046] Step S200, for the fast time-frequency domain signal By performing variable delay and constant delay in the slow time dimension, we obtain Parameterized instantaneous autocorrelation function ;

[0047] Step S300: Perform a variable-scale non-uniform Fourier transform on the relevant delay time to achieve coherent accumulation of the relevant delay time and eliminate t. m The influence of obtaining signals ;

[0048] Step S400, for the signal Obtain blind speed compensation ;

[0049] Step S500, the signal after blind speed compensation. Decoupling along the slow time using Fourier transform achieves energy accumulation for the relevant time;

[0050] Step S600: Apply inverse Fourier transform to the fast time frequency f to achieve coherent accumulation with respect to the distance frequency f;

[0051] Step S700: Obtain the target's range, velocity, and acceleration information based on the spectral peak positions in the coherent accumulation; compensate for the phase of the target's velocity and acceleration using the estimated velocity and acceleration; and perform coherent accumulation in the range-Doppler frequency domain.

[0052] t represents a fast-time variable, t m Let f represent the slow-time variable, and let f represent the frequency variable corresponding to the fast time.

[0053] In step S200, parameterize the instantaneous autocorrelation function. As shown in equation (1)

[0054] (1)

[0055] in, For delayed variables, The delay time is constant. B represents the amplitude of the signal, and B represents the bandwidth. t represents wavelength. m Let c represent the speed of light, and k represent the k-th target. Let a represent the initial distance to the k-th target. k Let v represent the initial acceleration of the k-th target.k Let f represent the initial velocity of the k-th target. c C represents the carrier frequency. TDPSIAF () indicates a cross term. In this embodiment, the amplitude changes accordingly after the signal changes, and A represents the amplitude.

[0056] The autocorrelation function represents how the similarity between the original sequence and its delayed sequence changes with the delay variable. The correlation delay represents the delay variable in the autocorrelation operation.

[0057] In equation (1), the exponential term is coupled with the distance frequency domain and the delay time variable, as well as the distance frequency domain and the slow time variable.

[0058] Equation (1) describes the echo signal as a single-frequency signal with respect to the delay variable. Therefore, in step S300, a variable-scale non-uniform Fourier transform is performed on the correlation delay time to achieve coherent accumulation of the correlation delay time, resulting in:

[0059] (2)

[0060] in, Indicates corresponding to The frequency variable; This is an introduced scaling factor used to adjust the PRVAD algorithm and the range and accuracy of velocity estimation. Through equation (2), the target echo information self-term energy is concentrated in the three-dimensional frequency-time acceleration space around ft. m Parallel planes superior.

[0061] In step S300, the parameterized distance-velocity-acceleration distribution algorithm uses de-line tuning processing to eliminate... The effect of the item, the process takes the following form.

[0062] (3)

[0063] The echo signal after demodulation was eliminated. The term contains only the distance frequency f and the slow time first-order term. .

[0064] Will and Substituting into equation (3), we can obtain

[0065] (4)

[0066] in, and Let X represent the autocorrelation radial baseband velocity and the autocorrelation radial velocity folding factor of the k-th target, respectively. This indicates the autocorrelation radial blind velocity, and * indicates the multiplication operator.

[0067] In step S400, the signal is searched using the velocity folding number method. That is, by performing blind speed compensation using formula (4), we obtain

[0068] (5)

[0069] In step S500, energy accumulation is achieved through a one-step variable-scale Fourier transform, as shown in formula (6).

[0070] (6)

[0071] in, Indicates corresponding to The frequency variable. At this time, the energy of the target echo signal is concentrated in the three-dimensional PFVAD space onto a straight line parallel to the f-axis. , superior.

[0072] In step S600, formula (7) is used to perform inverse Fourier transform on the fast time frequency f to achieve coherent accumulation with respect to the distance frequency f.

[0073] (7)

[0074] in, This represents the time variable corresponding to the distance frequency f, and rect() represents the gate function.

[0075] In step S700, the spectral peak positions of the gate function and the two Dirac functions are determined according to formula (7). The target range, velocity, and acceleration information are obtained. The estimated velocity and acceleration are used to compensate for these phase terms using formula (8), and coherent accumulation is achieved in the range-Doppler frequency domain.

[0076] (8)

[0077] in, Indicates the Doppler frequency. Indicates Fourier transform, This represents the inverse Fourier transform.

[0078] The methods for obtaining the estimated velocity and acceleration include: after the previous steps, processing the echo S... RF The coherent accumulation of () forms a spectral peak. The three-dimensional index is obtained by detecting the position of the spectral peak. The three-dimensional index is the frequency variable corresponding to the fast time, the slow time, and the frequency variable corresponding to the delay time. The velocity and acceleration parameter information is solved based on the correspondence between the three-dimensional index and the target distance, velocity, and acceleration.

[0079] Comparative Example

[0080] This embodiment improves the acceleration estimation range based on the original PRVAD algorithm. The acceleration estimation range and accuracy of the PRVAD algorithm are related to the pulse repetition frequency and the number of delay time points. Generally, in PD mode, high pulse repetition frequencies do not have the problem of insufficient acceleration measurement range, but in PC mode, low pulse repetition frequencies are likely to cause insufficient acceleration measurement range when detecting high-speed targets. However, the PRVAD algorithm does not consider this problem. This embodiment improves and updates the PRVAD algorithm by introducing a scaling factor in the delay time step, expanding the delay time frequency, thereby extending the acceleration measurement range of the PRVAD algorithm. This effectively solves the problem of missed detection of highly maneuvering targets when the target acceleration exceeds the radar system parameter measurement range.

[0081] The following is a simulation comparison before and after the algorithm improvement. The simulated radar system parameters are shown in the table below:

[0082] Table 1 Radar System Parameters

[0083]

[0084] Based on the radar system parameters provided in Table 1, the acceleration estimation range is: The calculated acceleration estimate range is (-38.4, 38.4) m / s². 2 The target's motion parameters are shown in Table 2, and the target's acceleration is 45 m / s². 2 The acceleration estimation range of the PRVAD algorithm is exceeded, so a scaling factor is introduced. The range of acceleration estimates is obtained as follows: The calculated value is (-76.8, 76.8) m / s. 2 .

[0085] Table 2 Target motion parameters

[0086]

[0087] Under the radar system parameters in Table 1, the acceleration estimation results before and after the PRVAD algorithm improvement are as follows when detecting moving targets in Table 2: Figure 2 As shown, the improved PRVAD algorithm expands the acceleration estimation range, significantly improves the signal-to-noise ratio at the target location, and increases the acceleration estimation result from 38.4 m / s². 2 To become an accurate 45m / s 2 .

[0088] The PRVAD algorithm, which incorporates a scale factor, estimates the acceleration to be 38.4 m / s². 2 The result of using this acceleration estimation to compensate for the original echo signal and coherently accumulating it is as follows: Figure 3As shown, the echo energy still diffuses. The acceleration estimated by the improved PRVA algorithm is 45 m / s², which, compared with the target motion parameters in Table 2, indicates that this is accurate acceleration information for the target. After compensation and coherent accumulation, the coherent accumulation result of the improved PRVAD algorithm is as follows: Figure 4 As shown, the echo energy is concentrated in a sharp spectral peak.

Claims

1. A method for extending the acceleration estimation range based on PRVAD, characterized in that, include: Step S100: Obtain the intermediate frequency echo data of K targets, and perform down-conversion mixing filtering and pulse compression to obtain the time domain echo signal. Perform fast-time Fourier transform on the time domain echo signal to obtain the fast-time frequency domain signal. Step S200: Let k=1, apply variable delay and constant delay to the fast time frequency domain signal in the slow time dimension, and obtain the autocorrelation of the fast time frequency domain signal; Step S300: Perform a variable-scale non-uniform Fourier transform on the correlation delay time in the autocorrelation of the fast time-frequency domain signal to achieve coherent accumulation of the correlation delay time; Step S400: De-line adjustment processing is used to eliminate the influence of slow time variables in the coherent accumulation of related delay times; Step S500: Perform blind speed compensation on the signal that eliminates the influence of slow time variables; Step S600: Perform Fourier transform on the signal after blind speed compensation along the slow time to decouple it and achieve energy accumulation for the relevant time. Step S700: Apply inverse Fourier transform to the fast time frequency to achieve coherent accumulation of the distance frequency, and obtain the target distance, velocity, and acceleration information; Step S800: Compensate the obtained velocity and acceleration with the estimated velocity and acceleration; If k < K, proceed to step S200, k = k + 1, and replace the target distance, velocity, and acceleration in the fast time-frequency domain signal; if k = K, output the compensated acceleration. In step S200, the autocorrelation function R of the fast time-frequency domain signal is obtained. TDPSIAF (f,t m ,τ m As shown in formula (1) Where, τ m h is a time-delay variable. m Let A be a constant delay time. 2k Let B represent the amplitude of the signal, λ represent the bandwidth, and t represent the wavelength. m Let c represent the speed of light, k represent the k-th target, and R represent the slow-time variable. 0k Let a represent the initial distance to the k-th target. k Let v represent the initial acceleration of the k-th target. k Let f represent the initial velocity of the k-th target. c C represents the carrier frequency. TDPSIAF () indicates a cross term. In step S300, a variable-scale non-uniform Fourier transform is performed on the relevant delay time according to equation (2) to achieve coherent accumulation of the relevant delay time. in, This indicates the correspondence to τ. m The frequency variable, ξ, is the introduced scaling factor, A 3k C represents the amplitude of the transformed signal. PFTAD () represents the cross term, and δ() is the Dirac function.

2. The method according to claim 1, characterized in that, Eliminate the influence of slow time variables according to equation (3) Among them, A 4k C represents the amplitude of the transformed signal. DP () indicates a cross term.

3. The method according to claim 2, wherein the blind speed compensation process in step S500 includes: v k =v 0k +n k v b and exp(-j2πn k f p t m Substituting 1 into equation (3), we get Among them, v 0k and n k Let v represent the autocorrelation radial baseband velocity and the autocorrelation radial velocity folding factor of the k-th target, respectively. b f represents the autocorrelation radial blind velocity. p Indicates the pulse repetition period. The mismatch term representing the velocity folding factor; The signal is searched using the velocity folding number method. Perform blind speed compensation, get 4. The method according to claim 3, characterized in that, Energy accumulation is achieved through a one-step variable-scale Fourier transform, as shown in formula (6). in, This indicates that the corresponding to t m The frequency variable, The mismatch term representing the velocity folding factor, Indicates the intersection of terms.

5. The method according to claim 4, characterized in that, In step S700, formula (7) is used to perform inverse Fourier transform on the fast time frequency f to achieve coherent accumulation with respect to the distance frequency f. Among them, t n ' represents the time variable corresponding to the distance frequency f. The mismatch term representing the velocity folding factor, Indicates the intersection of terms.

6. The method according to claim 5, characterized in that, Compensation for velocity and acceleration is performed according to equation (8). Among them, f d Indicates the Doppler frequency. Indicates Fourier transform, This represents the inverse Fourier transform.

Images