A motion parameter estimation method for maneuvering faint targets based on Stepped-DGRFT

Abstract

The present invention relates to the field of maneuvering weak target detection using a de-skewed stepped-frequency radar, and more specifically to a maneuvering weak target motion parameter estimation and coherent accumulation method based on Stepped-DGRFT. The present invention proposes a maneuvering weak target motion parameter estimation method based on Stepped-DGRFT, which can accurately estimate the motion parameters of maneuvering weak targets using a de-skewed stepped-frequency radar. Combined with motion parameter compensation and a one-dimensional target range profile focusing method, the system can improve target detection capabilities. Compared with the DGRFT method, Stepped-DGRFT can effectively improve the target motion parameter estimation and imaging capabilities of a de-skewed stepped-frequency radar for maneuvering weak targets.

Description

Technical Field

[0001] The present invention relates to the field of maneuvering weak target detection using a de-skewed stepped-frequency radar, and more specifically, to a maneuvering weak target motion parameter estimation and coherent accumulation method based on Stepped-DGRFT. The present invention proposes a maneuvering weak target motion parameter estimation method based on Stepped-DGRFT, which can accurately estimate the motion parameters of maneuvering weak targets using a de-skewed stepped-frequency radar. Combined with motion parameter compensation and a one-dimensional target range profile focusing method, the system can improve target detection capabilities. Compared to the DGRFT method, Stepped-DGRFT can effectively improve the target motion parameter estimation and imaging capabilities of a de-skewed stepped-frequency radar for maneuvering weak targets. Background Art

[0002] In recent years, stealth-capable aircraft have continued to emerge. These small RCS targets, such as stealth aircraft, pose a significant threat and challenge to radar detection. Maintaining high-precision measurement of the motion parameters of maneuvering targets with small RCS is a pressing challenge for modern radars. To achieve high-resolution imaging and accurate estimation of motion parameters of maneuvering targets with small RCS, radars must utilize broadband or even ultra-wideband signals. This allows them to utilize broadband echoes to obtain more and richer target information for further target classification and identification. While it is theoretically possible to directly transmit and receive wide-bandwidth LFM signals and combine them with de-skewing processing to achieve high-resolution target imaging, the limitations of the system's RF unit design and front-end sampling rate make direct de-skewing of ultra-wideband signals difficult. Furthermore, the large bandwidth can easily introduce amplitude and phase distortion, hindering refined processing such as amplitude, phase, and velocity compensation, thus affecting target imaging. Frequency-stepped signals, however, can achieve wider-bandwidth composite imaging through coherent processing of multiple frames, while maintaining the smaller bandwidth of the sub-pulses. Furthermore, multiple sub-pulses can be individually processed for amplitude, phase, and velocity compensation, further improving imaging quality. Therefore, it is possible to combine sub-pulse broadband de-skewing with frequency stepping to achieve a de-skewing, frequency-stepped, wide-bandwidth signal (Jiang Z, Wang J, Song Q, et al. Obstacle sensing using dechirp SAR[C] / / 2016IEEE 13th International Conference on Signal Processing (ICSP). IEEE, 2016: 1456-1460. Kelly SI, Davies ME. RFI suppression and sparse image formation for UWB SAR[C] / / 2013 14th International Radar Symposium (IRS). IEEE, 2013, 2: 655-660.). Specifically, a broadband linear frequency modulation signal is transmitted within a sub-pulse and its echo is de-skewed, while frequency stepping is performed between pulses. Combined with coherent processing methods such as spectral splicing and time-domain splicing, as well as amplitude, phase, and velocity compensation, target imaging over a wide bandwidth can be achieved.

[0003] In the de-slanted frequency stepped radar system, in order to obtain a one-dimensional high-resolution range profile (HRRP) with good focusing performance of a maneuvering weak target through synthetic broadband processing to achieve target detection and identification, it is first necessary to accurately estimate the target's motion parameters and use the parameter estimation results to perform motion compensation and coherent accumulation on the target echo. However, the existing de-slanted maneuvering weak target motion parameter estimation method is difficult to meet this requirement. The DGRFT algorithm (You P, Ding Z, Liu S, et al. Dechirp-receiving radar target detection based on generalized Radon-Fourier transform [J]. IET Radar, Sonar & Navigation, 2021, 15 (9): 1096-1111.) is proposed based on a constant carrier frequency radar system. Although the de-slanted frequency stepped signal can also be processed as a single carrier frequency signal after synthetic broadband processing, the motion parameter estimation required for target motion compensation during the synthetic broadband processing process needs to be obtained through a coherent accumulation algorithm. This is obviously inconsistent with the construction steps of the ideal HRRP. The present invention proposes a modified DGRFT method (i.e., Stepped-DGRFT) to achieve long-term coherent accumulation and accurate estimation of motion parameters of maneuvering weak targets before synthetic broadband processing. Combined with motion compensation and synthetic broadband processing, the HRRP focusing effect of maneuvering weak targets is effectively improved.

[0004] Therefore, based on the de-sloped frequency stepped radar system, studying a motion parameter estimation method for maneuvering weak targets has important practical significance and application value. Summary of the Invention

[0005] The technical problem solved by the present invention is: to overcome the shortcomings of the existing technology and propose a motion parameter estimation method for a maneuvering faint target based on Stepped-DGRFT. The method first de-skews the target baseband echo and compensates the phase term, then searches for the target motion parameters in the parameter space, and finally performs motion compensation on the maneuvering faint target in the time domain according to the target motion parameter estimation results to achieve the effect of one-dimensional range image focusing.

[0006] To achieve the above object, the technical solution of the present invention is:

[0007] A method for estimating motion parameters of a maneuvering faint target based on Stepped-DGRFT includes the following steps:

[0008] Step 1: down-convert the target radar echo to obtain a target baseband echo, and perform de-skewing processing on the obtained target baseband echo to obtain a de-skewing echo signal;

[0009] Step 2: construct a phase compensation factor, and use the constructed phase compensation factor to compensate the phase term of the de-skewed echo signal obtained in step 1;

[0010] Step 3: Perform a traversal search on the target motion parameters in the phase compensation factors constructed in Step 2 to obtain an accumulation result matrix. If and only if the searched target motion parameters are equal to the actual target motion parameters, the accumulation result matrix will have a peak. The target motion parameters corresponding to the peak are the accurate estimation results of the target motion parameters.

[0011] Step 4, evaluate the accuracy of the estimated value of the target motion parameter, the method is: use the accurate estimation result of the target motion parameter obtained in step 3 to perform motion compensation on the de-slanted echo signal obtained in step 1 in the time domain, synthesize the time domain target de-slanted echo after motion compensation into a wideband through the time domain splicing method to obtain a one-dimensional range image of each frame signal, because the target is a weak target with a low signal-to-noise ratio characteristic, the one-dimensional range image of each frame signal is summed in the slow time dimension to obtain a one-dimensional range image after coherent accumulation, thereby realizing the focusing of the one-dimensional range image of the maneuvering weak target, and obtaining the accuracy of the target motion parameter estimation value based on the error between the one-dimensional distance of the peak of the focusing result and the target distance set by the simulation parameters.

[0012] In the step 1, the target radar echo is

[0013]

[0014] Where n = (m-1)N + h is the sub-pulse number, m is the frame number, m = 1, 2, ..., M. N is the number of sub-pulses in each frame signal, and h is the sub-pulse number in the frame. p Indicates fast time, T p is the pulse width, and defines is the frequency modulation slope. f h =f0+hΔf is the carrier frequency of each sub-pulse, h=0,1,...,N-1, f0 is the initial carrier frequency, Δf is the frequency step interval, and the pulse repetition time (PRT) is T r , accumulation time T = MNT r .

[0015] In the step 1, the target baseband echo is:

[0016]

[0017] Among them, K ris the amplitude factor of the target echo. For simplicity, it is assumed that it remains unchanged between pulses within the same CPI, and c is the speed of light. T is the target motion parameter set of each order, α T =[α0,α1,α2,…,α K-1 ], α0, α1, α2, ... respectively represent the target's initial radial distance, initial velocity, initial acceleration, and so on. The instantaneous radial distance of the target can be expressed as:

[0018]

[0019] In the step 1, the de-slanted echo signal s r (n,t p )for:

[0020]

[0021] The first term of the phase is the instantaneous radial distance of the target The second term is the phase in the range after deslanting. The changing azimuth phase is the Doppler phase history; the third term is the residual video phase (RVP). ref (t) is the de-skew reference signal:

[0022]

[0023] Among them, R ref is the reference distance. The de-skewing process is the conjugate multiplication of the echo signal and the reference signal in the time domain. The time-frequency relationship of the de-skewing frequency step signal is as follows: Figure 1 shown.

[0024] In the step 2, the constructed phase compensation factor is:

[0025]

[0026] In the step 2, the constructed phase compensation factor includes target motion parameters, which include the distance α0 between the initial position of the target and the radar, the initial velocity of the target α1, and the initial acceleration of the target α2.

[0027] In step 3, the accumulation result matrix G is:

[0028]

[0029] In the fourth step, when the target is assumed to move at a uniform radial speed, Stepped-DGRFT degenerates into Stepped-DRFT, and the de-skewed echo signal s after motion compensation is obtained. rm(n,t p )for:

[0030]

[0031] in, is the target initial distance and initial speed estimation result in step 3.

[0032] Beneficial effects

[0033] This paper proposes a method for motion parameter estimation and coherent accumulation of maneuvering faint targets based on Stepped-DGRFT. This method enables accurate motion parameter estimation of maneuvering faint targets using a de-skewed stepped-frequency radar. Combined with motion parameter compensation and a one-dimensional range profile focusing method, this method improves the system's target detection capabilities. Compared to the DGRFT method, Stepped-DGRFT effectively enhances the motion parameter estimation and imaging capabilities of de-skewed stepped-frequency radars for maneuvering faint targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is the time-frequency relationship of the de-ramped frequency-stepped signal;

[0035] Figure 2 This is the target motion parameter estimation result based on the Stepped-DRFT algorithm;

[0036] Figure 3 This is the HRRP focusing result based on the Stepped-DRFT algorithm;

[0037] Figure 4 This is the HRRP focusing result based on the Stepped-DRFT algorithm (partial magnification);

[0038] Figure 5 This is the target motion parameter estimation result based on the traditional DRFT algorithm;

[0039] Figure 6 This is the HRRP focusing result based on the traditional DRFT algorithm;

[0040] Figure 7 The target motion parameter estimation result (range-velocity plane) based on the Stepped-DRFT algorithm;

[0041] Figure 8 The target motion parameter estimation result (acceleration-range plane) based on the Stepped-DRFT algorithm;

[0042] Figure 9 This is the HRRP focusing result based on the Stepped-DRFT algorithm;

[0043] Figure 10 This is the HRRP focusing result based on the Stepped-DRFT algorithm (partial magnification);

[0044] Figure 11 The target motion parameter estimation result (range-velocity plane) based on the traditional DRFT algorithm;

[0045] Figure 12 This is the target motion parameter estimation result (acceleration-range plane) based on the traditional DRFT algorithm;

[0046] Figure 13 This is the HRRP focusing result based on the traditional DRFT algorithm. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0048] The specific steps of a motion parameter estimation method for maneuvering faint targets based on Stepped-DGRFT are as follows:

[0049] Step 1: De-bevel.

[0050] The de-sloped frequency stepped radar transmission signal can be expressed as

[0051]

[0052] Where n = (m-1)N + h is the sub-pulse number, m is the frame number, m = 1, 2, ..., M. N is the number of sub-pulses in each frame signal, and h is the sub-pulse number in the frame. p Indicates fast time, T p is the pulse width, and defines is the frequency modulation slope. f h =f0+hΔf is the carrier frequency of each sub-pulse, h=0,1,...,N-1, f0 is the initial carrier frequency, Δf is the frequency step interval, and the pulse repetition period is T r , accumulation time T = MNT r .

[0053] Considering a single point target scenario, assuming that the target motion parameters of each order are represented by a set α T =[α0,α1,α2,…,α K-1 ], where α0, α1, α2, ... respectively represent the target's initial radial distance, initial velocity, initial acceleration, and so on. The instantaneous radial distance of the target can be expressed as:

[0054]

[0055] Based on the stop-and-go model, the target echo after down-conversion to baseband can be obtained as:

[0056]

[0057] Among them, K r is the amplitude factor of the target echo. For simplicity, it is assumed to remain unchanged between pulses within the same CPI, and c is the speed of light.

[0058] The de-skew reference signal is:

[0059]

[0060] Among them, R ref is the reference distance. The down-conversion process is the conjugate multiplication of the echo signal and the reference signal in the time domain. The de-skewing echo signal can be expressed as:

[0061]

[0062] The first term of the phase is the instantaneous radial distance of the target The second term is the phase in the range after deslanting. The changing azimuth phase is the Doppler phase history. The third item is the residual video phase (RVP).

[0063] Step 2: Phase compensation.

[0064] In order to achieve effective coherent accumulation of multi-frame signals, the phase term in the de-slanted echo signal needs to be To compensate, the corresponding filter group, i.e. the phase compensation factor, is as follows:

[0065]

[0066] Step 3: Parameter estimation.

[0067] Through steps one and two, we can get the coherent integration algorithm for detecting maneuvering weak targets with de-slanted frequency stepped signals.

[0068]

[0069] The echo signal after de-skewing s r (n,t p ) After compensation by the filter bank h(n,α), a one-dimensional range image of the target is obtained by performing FFT in the time domain. After compensation by the filter bank, the peaks of the one-dimensional range images of all de-skewed sub-pulses appear at the reference distance. By traversing the target motion parameters, the coherent accumulation output of the target echo can be obtained. If and only if the search parameter set α satisfies α=αT When the de-sloped frequency step signal is used to detect the maneuvering weak target, the output corresponding to the reference distance after the coherent superposition of MN one-dimensional range images in the coherent accumulation algorithm in step 3 is the largest, that is, the full coherent accumulation of the target is achieved. Each set of parameters in the search parameter set corresponds to the value at the reference distance after accumulation. By performing peak detection on the result matrix, an accurate estimation result of the target motion parameter can be obtained from the parameter search space. Taking into account the carrier frequency hopping characteristics between pulses of the de-sloped frequency step signal, the present invention refers to the DGRFT algorithm defined in step 3 for the de-sloped frequency step signal as a Stepped-DGRFT algorithm, which is different from the classic DGRFT.

[0070] For most application scenarios, the radial velocity of the target during the accumulation time can be considered constant. In this case, only the initial radial distance R T and radial velocity v T Then Stepped-DGRFT can be simplified to Stepped-DRFT, that is

[0071]

[0072] in

[0073]

[0074] is the degraded filter bank, r is the initial radial distance of the target, and v is the radial velocity.

[0075] Step 4: Motion compensation and focusing.

[0076] Correct motion compensation is a prerequisite for one-dimensional range image focusing. According to the motion parameter estimation results of Stepped-DRFT, motion compensation is performed on the maneuvering weak target in the time domain.

[0077]

[0078] After motion compensation based on Stepped-DRFT, a decimation-based time-domain splicing method is used to synthesize the wideband signal to obtain a one-dimensional range image for each frame. Because the target is faint and has a low signal-to-noise ratio, the one-dimensional range image of M frames is summed in the slow time dimension to obtain a coherently accumulated one-dimensional range image, thus achieving focusing of the one-dimensional range image of the maneuvering faint target.

[0079] The present invention provides the following examples to illustrate the inventive method:

[0080] The verification conditions of the implementation example provided by the present invention under the condition of uniform target motion are shown in Table 1:

[0081] Table 1 Implementation example verification conditions

[0082]

[0083] For Stepped-DRFT and traditional DGRFT, a unified velocity search range is set to 0 m / s to 200 m / s, and the velocity direction is defined as positive when the target moves away from the radar. Figure 2 The target motion parameter estimation results obtained using the Stepped-DRFT algorithm are given. The target speed is estimated to be 50m / s with an error of 0m / s. Figure 3 and Figure 4 The target HRRP is obtained after motion compensation, time-domain splicing and broadband synthesis, and coherent integration. The proposed algorithm can accurately estimate and compensate target parameters. In the target HRRP obtained after coherent integration, the two scattering points of the target are accurately located, with an amplitude difference of approximately 0.07 dB, indicating good HRRP focusing performance.

[0084] When using traditional DRFT, it is necessary to first obtain the target HRRP through synthetic wideband processing without motion compensation, and then use DRFT to perform long-term coherent accumulation of the target HRRP. Figure 5 The target motion parameter estimation results obtained using the DRFT algorithm are given. The target velocity is estimated to be 44.32m / s with an error of 5.68m / s. Figure 6 The target HRRP is obtained after motion compensation, time-domain stitching and wideband synthesis, and coherent integration. As can be seen, when using the traditional DRFT algorithm, the two scattering points of the target are submerged in the noise, and the two HRRP peaks appear at 300.061714 km and 299.983333 km, respectively. The one-dimensional range image cannot be correctly focused.

[0085] To conduct a detailed performance comparison between Stepped-DRFT and traditional DRFT algorithms under target maneuvers, further simulations were conducted with the target in uniformly accelerated motion, simulating a target maneuvering scenario. The target consisted of two scattering points of equal intensity, initially at radial distances of 300,000.000m and 300,000.625m. During the entire accumulation time, the target spanned 12 range cells, exhibiting a significant ARU phenomenon.

[0086] The verification conditions for the implementation example of the target uniform acceleration maneuvering motion provided by the present invention are shown in Table 2:

[0087] Table 2 Implementation example verification conditions

[0088]

[0089] For Stepped-DRFT and traditional DGRFT, the speed search range is set to 0m / s to 200m / s, and the speed direction is positive when the target moves away from the radar, and the acceleration search range is -10m / s. 2 ~10m / s 2 . Figure 7 and Figure 8 The target motion parameter estimation results obtained using the Stepped-DRFT algorithm are given. The target initial velocity is estimated to be 50m / s with an error of 0m / s; the target acceleration is estimated to be 3.6m / s 2 , the error is 0m / s 2 . Figure 9 and Figure 10 The target HRRP is obtained after motion compensation, time-domain splicing and broadband synthesis, and coherent integration. The proposed algorithm can accurately estimate and compensate target parameters. In the target HRRP obtained after coherent integration, the two scattering points of the target are accurately located, with an amplitude difference of approximately 0.33 dB, indicating good HRRP focusing performance.

[0090] Figure 11 and Figure 12 The target motion parameter estimation results obtained using the DRFT algorithm are given. The target initial velocity is estimated to be 18.97m / s with an error of 31.03m / s; the target acceleration is estimated to be 9.10m / s 2 , the error is 4.10m / s 2 . Figure 13 The target HRRP is obtained after motion compensation, time-domain stitching and wideband synthesis, and coherent integration. As can be seen, when using the traditional DRFT algorithm, the two scattering points of the target are submerged in the noise, and the two HRRP peaks appear at 300.090351 km and 300.115840 km, respectively, and the one-dimensional range image cannot be correctly focused.

[0091] From the above results, it can be seen that after using the method proposed in the present invention for parameter estimation and motion compensation, accurate estimation of target motion parameters and focusing of one-dimensional range profiles in weak target maneuvering scenarios under the de-slanted frequency stepped radar system are achieved. Compared with the existing method DRFT, the superiority of the method of the present invention is demonstrated. In summary, the above is only a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for estimating motion parameters of maneuvering faint targets based on Stepped-DGRFT, characterized by The steps include: Step 1: down-convert the target radar echo to obtain a target baseband echo, and perform de-skewing processing on the obtained target baseband echo to obtain a de-skewing echo signal; Step 2: construct a phase compensation factor, and use the constructed phase compensation factor to compensate the phase term of the de-skewed echo signal obtained in step 1; Step 3: Perform a traversal search on the target motion parameters in the phase compensation factors constructed in Step 2 to obtain an accumulation result matrix. If and only if the searched target motion parameters are equal to the actual target motion parameters, the accumulation result matrix will have a peak. The target motion parameters corresponding to the peak are the accurate estimation results of the target motion parameters. Step 4: Evaluate the accuracy of the estimated target motion parameters by using the accurate target motion parameter estimation results obtained in step 3 to perform motion compensation in the time domain on the de-skewed echo signal obtained in step 1. The motion-compensated time-domain de-skewed echo is synthesized into a wideband using a time-domain splicing method to obtain a one-dimensional range image for each frame of the signal. The one-dimensional range image of each frame signal is summed in the slow time dimension to obtain the one-dimensional range image after coherent accumulation. The accuracy of the target motion parameter estimation value is obtained based on the error between the one-dimensional distance of the peak of the focusing result and the target distance set by the simulation parameters. In the step 2, the constructed phase compensation factor is: Among them, h(n,α) is the constructed phase compensation factor, c is the speed of light, k is the frequency modulation slope, is the instantaneous radial distance of the target, n is the sub-pulse number, α is the set of target motion parameters of each order, R ref is the reference distance, t is the fast time, f h is the carrier frequency of each sub-pulse.

2. The method for estimating motion parameters of a maneuvering faint target based on Stepped-DGRFT according to claim 1, characterized in that: In the step 1, the target radar echo is Where n = (m-1)N + h is the sub-pulse number, m is the frame number, m = 1, 2, ..., M; N is the number of sub-pulses in each frame signal, h is the sub-pulse number in the frame; t p Indicates fast time, T p is the pulse width, and defines is the frequency modulation slope; f h =f0+h△f is the carrier frequency of each sub-pulse, h=0,1,...,N-1, f0 is the initial carrier frequency, △f is the frequency step interval, and the pulse repetition time (PRT) is T r , accumulation time T = MNT r .

3. The method for estimating motion parameters of a maneuvering faint target based on Stepped-DGRFT according to claim 2, characterized in that: In the step 1, the target baseband echo is: Among them, K r is the amplitude factor of the target echo. For simplicity, it is assumed that it remains unchanged between pulses within the same CPI. c is the speed of light. α T is the set of target motion parameters of each order, α T =[α0,α1,α2,…,α K-1 ], α0, α1, α2, ... respectively represent the target's initial radial distance, initial velocity, initial acceleration, and so on; the instantaneous radial distance of the target can be expressed as:

4. The method for estimating motion parameters of a maneuvering faint target based on Stepped-DGRFT according to any one of claims 1 to 3, characterized in that: In the step 1, the de-slanted echo signal s r (n,t p )for: The first term of the phase is the instantaneous radial distance of the target The second term is the phase in the range after deslanting. The changing azimuth phase, that is, the Doppler phase process; the third term is the residual video phase (RVP) term; where s ref (t) is the de-skew reference signal: Among them, R ref is the reference distance; the de-skewing process is the conjugate multiplication of the echo signal and the reference signal in the time domain.

5. The method for estimating motion parameters of a maneuvering faint target based on Stepped-DGRFT according to claim 1, characterized in that: In the step 2, the constructed phase compensation factor includes target motion parameters, which include the distance α0 between the initial position of the target and the radar, the initial velocity of the target α1, and the initial acceleration of the target α2.

6. The method for estimating motion parameters of a maneuvering faint target based on Stepped-DGRFT according to claim 1, characterized in that: In step 3, the accumulation result matrix G is:

7. The method for estimating motion parameters of a maneuvering faint target based on Stepped-DGRFT according to claim 1, characterized in that: In the fourth step, when the target is assumed to move at a uniform radial speed, Stepped-DGRFT degenerates into Stepped-DRFT, and the de-skewed echo signal s after motion compensation is obtained. rm (n,t p )for: in, is the target initial distance and initial speed estimation result in step 3.

8. The method for estimating motion parameters of a maneuvering faint target based on Stepped-DGRFT according to claim 1, characterized in that: The accuracy of the estimated value of the target motion parameter is evaluated by using the accurate estimation result of the target motion parameter obtained in step three to perform motion compensation on the de-slanted echo signal obtained in step one in the time domain. The de-slanted echo of the time domain target after motion compensation is synthesized into a broadband through the time domain splicing method to obtain a one-dimensional range image of each frame signal. Since the target is a weak target with a low signal-to-noise ratio characteristic, the one-dimensional range image of each frame signal is summed in the slow time dimension to obtain a one-dimensional range image after coherent accumulation, thereby achieving the focusing of the one-dimensional range image of the maneuvering weak target, and the accuracy of the target motion parameter estimation value is obtained based on the error between the one-dimensional distance of the peak of the focusing result and the target distance set by the simulation parameters.

Images