A distributed aperture radar deep space target high-resolution imaging algorithm based on wideband adaptive coherent synthesis

Abstract

This invention discloses a high-resolution imaging algorithm for deep-space targets using distributed aperture radar based on broadband adaptive coherent synthesis. First, adjacent carrier frequency signals are transmitted by different radar units. After reception, the carrier frequencies are separated by filtering and pulse compression is performed. Second, a coarse search for time delay error and a search for frequency error are conducted using a search algorithm based on a hybrid mutual ambiguity function. Third, a fine search for time delay error and phase error is conducted using a search algorithm based on distributed coherent generalized Rayleigh-Fourier transform. Finally, after compensating for the searched errors, the signals from each radar unit are coherently synthesized into an equivalent large-bandwidth signal, which is then used for high-resolution imaging using a delayed Doppler algorithm.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing, and in particular to a high-resolution imaging algorithm for deep-space targets using distributed aperture radar based on broadband adaptive coherent synthesis. Background Technology

[0002] Deep space radar plays a crucial role in space exploration, planetary defense, and the characterization of small celestial bodies. As deep space exploration missions increasingly demand higher spatial resolution and observation ranges, radar systems need to achieve high gain and high-resolution imaging under extremely weak echo conditions. However, for deep space targets, the extremely long propagation distances lead to severe path loss. Furthermore, limitations in transmitter design, sampling rate, and data transmission hardware restrict traditional monostation radars from generating and processing broadband signals over extremely long propagation distances, thus hindering further improvements in imaging resolution.

[0003] In order to overcome the physical bottleneck of monostation radar, distributed frequency stepping radar system has become an effective alternative [1]. This system uses multiple spatially separated radar units to transmit and receive narrowband signals of different carrier frequencies and synthesize an equivalent broadband signal, thereby greatly improving resolution and processing gain without requiring a single radar to have instantaneous broadband capability. However, in practical applications, due to the spatial separation of radar units and non-ideal factors such as oscillator instability, ephemeris error, and channel delay, there are serious delay mismatches and phase shifts between the receiving channels [2]. These factors are coupled with each other, resulting in envelope mismatch and signal decorrelation problems in the synthesized broadband signal, which seriously reduces the coherent accumulation performance and actual imaging resolution. Existing broadband synthesis methods are mostly designed for monostation radar and are difficult to apply directly to distributed frequency stepping radar systems in scenarios with extremely low signal-to-noise ratios.

[0004] [1] Kuang Yunlian, Lu Jun, Hu Guangmin. Clutter spectrum analysis of airborne distributed coherent MIMO radar [J]. Journal of China Academy of Electronics Science, 2014, 9(1):5.DOI:10.3969 / j.issn.1673-5692.2014.01.011. [2] Brown DR , Poor HV .Time-Slotted Round-Trip CarrierSynchronization for Distributed Beamforming[J].IEEE Transactions on SignalProcessing, 2008.DOI:10.1109 / TSP.2008.927073. Summary of the Invention

[0005] In view of this, the present invention provides a high-resolution imaging algorithm for deep-space targets using distributed aperture radar based on broadband adaptive coherent synthesis. This method uses celestial bodies as illumination targets and aims to obtain high-resolution two-dimensional images of celestial bodies.

[0006] This method estimates and compensates for the errors between radar elements using an optimization algorithm based on a hybrid mutual fuzzy function and a distributed coherent generalized Raydon-Fourier transform, and employs a cascaded strategy to achieve fast and efficient search.

[0007] The technical solution of this invention is as follows: A high-resolution imaging algorithm for deep-space targets using distributed aperture radar based on broadband adaptive coherent synthesis includes: Step 1: Select the signal transmitted by a certain radar unit as a reference. For adjacent carrier frequency signals transmitted by different radar units, after receiving them by the same antenna, perform down-conversion to baseband and pulse compression. Step 2: Using the reference signal selected in Step 1 as a benchmark, divide the time delay error search range and the fine frequency error search range. Accumulate the signal to be calibrated and the reference signal based on the hybrid mutual ambiguity function. When the accumulated value is the maximum, obtain the corresponding time delay and frequency error search results. Step 3: Based on the search results based on the hybrid mutual ambiguity function, after compensating for the frequency error of each radar unit signal, divide the search range for time delay and phase error, and perform signal energy accumulation based on distributed coherent generalized Raydon-Fourier transform. When the accumulated value is at its maximum, the accurate search results for time delay and phase error are obtained. Step 4: Based on the searched time delay, frequency, and phase error, the signals of each radar unit are compensated and then coherently synthesized to obtain an equivalent large bandwidth signal. Then, the classic delayed Doppler imaging algorithm is used to obtain high-resolution imaging results.

[0008] Furthermore, the pulse-compressed signal expression in step one (taking the i-th radar unit transmitting and the 1st radar unit receiving as a reference unit as an example) is as follows: (1) in, and They represent fast time and slow time, respectively. Indicates signal amplitude. Represents the envelope function. Indicates the frequency modulation slope. Indicates frequency error. Indicates the actual delay. Indicates the actual carrier frequency. Indicates the initial phase.

[0009] Furthermore, the cumulative expression based on the hybrid mutual fuzzy function in step two is as follows: (2) in, Indicates the number of sub-apertures. and These represent the time delay and frequency error to be searched, respectively.

[0010] The time delay and frequency error corresponding to the maximum accumulated value are: (3) Furthermore, the accumulation expression based on the distributed coherent generalized Raydon-Fourier transform in step three is as follows: (4) in, This represents the number of pulses within a sub-aperture. and These represent the time delay and phase error to be searched, respectively. This represents the signal after compensating for frequency errors.

[0011] The search results for the time delay and phase error corresponding to the maximum accumulated value are as follows: (5) Furthermore, the coherent synthesis expression after error compensation for each unit signal in step four is as follows: (6) in, This indicates the peak position after signal accumulation.

[0012] Beneficial effects: 1. This invention proposes a distributed step-frequency imaging mode, which improves the resolution of the imaging results by coherently combining the narrowband signals of different radar units into a broadband signal. The resolution improvement ratio is proportional to the number of radar units. 2. This invention proposes an error search algorithm based on a hybrid mutual fuzzy function, which can quickly achieve coarse estimation of time delay error and fine estimation of frequency error between transmitted signals of different radar units, with a frequency error estimation accuracy of up to 98.9%. 3. This invention proposes an error search algorithm based on distributed coherent generalized Raydon-Fourier transform, which can quickly and accurately estimate the time delay and phase error between the transmitted signals of different radar units. The time delay error estimation accuracy can reach 96.3%, and the phase error estimation accuracy can reach 88.9%. Attached Figure Description

[0013] Figure 1 A schematic diagram of the process of this invention; Figure 2 Distributed step-frequency radar imaging model; Figure 3Simulation model schematic diagram: (a) Optical model; (b) Point cloud model; Figure 4 Search results for time-frequency errors based on hybrid mutual fuzzy functions: (a) Search results for time delay-frequency errors; (b) Search results for one-dimensional time delay errors; (c) Search results for one-dimensional frequency errors; Figure 5 Search results for time-phase error based on distributed coherent generalized Raydon-Fourier transform: (a) Search results for time delay-phase error; (b) Search results for one-dimensional time delay error; (c) Search results for one-dimensional phase error. Figure 6 Computer simulation results of two-dimensional imaging: (a) Single-station radar imaging results; (b) Multi-station radar coherent composite imaging results before error compensation; (c) Multi-station radar coherent composite imaging results after error compensation; Figure 7 Search results for asteroid measured data errors: (a) Time delay error search results; (b) Frequency error search results; (c) Phase error search results; Figure 8 Two-dimensional imaging results of asteroid measured data: (a) Single-station radar asteroid imaging results; (b) Multi-station radar coherent synthesis asteroid imaging results before error compensation; (c) Multi-station radar coherent synthesis asteroid imaging results after error compensation. Detailed Implementation

[0014] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0015] This invention discloses a high-resolution imaging algorithm for deep-space targets using distributed aperture radar based on broadband adaptive coherent synthesis. Taking celestial objects as targets, the algorithm first receives signals transmitted by different radar units using the same antenna, then down-converts them to baseband and performs pulse compression. Next, a specific radar unit is selected as a reference, and a time delay error search range and a fine-grained frequency error search range are defined. Time delay and frequency errors are searched based on the accumulation results of a hybrid cross-ambiguity function (HCAF). Then, the searched frequency errors are compensated, and a fine-grained time delay error search grid and a phase error search grid are defined. Signal energy accumulation based on a distributed coherent generalized Radon-Fourier transform (DC-GRFT) is performed. The most accurate time delay and phase error search results are obtained when the accumulated value is maximized. Finally, after compensating for the searched time delay, frequency, and phase errors, the signals from each radar unit are coherently synthesized and the classic delayed Doppler imaging algorithm is applied to obtain a high-resolution imaging result.

[0016] Appendix Figure 1 This is a schematic diagram of the process of the present invention. The method specifically includes the following four steps: Step 1: Select the signal transmitted by a specific radar unit as a reference. For adjacent carrier frequency signals transmitted by different radar units, after receiving them with the same antenna, perform down-conversion to baseband and pulse compression: In a distributed step-frequency radar system, different radar elements transmit signals at adjacent carrier frequencies, as shown in the attached diagram. Figure 2 As shown. After being received by the receiving antenna, signals of different carrier frequencies are separated by filtering and then pulse compressed. Taking the i-th radar unit transmitting and the 1st radar unit receiving as a reference unit as an example, the expression of the pulse-compressed signal is: (1) in, and They represent fast time and slow time, respectively. Indicates signal amplitude. Represents the envelope function. Indicates the frequency modulation slope. Indicates frequency error. Indicates the actual delay. Indicates the actual carrier frequency. Indicates the initial phase.

[0017] Step 2: Using the reference signal selected in Step 1 as a benchmark, divide the time delay error search range and the fine frequency error search range. Accumulate the signal to be calibrated and the reference signal based on the hybrid mutual ambiguity function. When the accumulated value is maximum, obtain the corresponding time delay and frequency error search results: The pulse compression signal obtained in step one is then accumulated based on a hybrid mutual ambiguity function. Let the total number of observed pulses be... Divide it evenly into Each coherent processing interval contains _ coherent processing intervals_. One pulse. Divide the time delay error search grid into a coarse one and a fine one into a frequency error search grid, and let the time delay error search value be... The frequency offset error search value is Construct a joint compensation operator for the path clipping time to be calibrated: (2) The operator is applied to the signal to be calibrated, performing coherent accumulation within a single coherent processing interval, followed by... Non-coherent superposition is performed between each coherent processing interval. The expression for the hybrid mutual fuzzy function with two-dimensional parameter optimization is then: (3) Therefore, by extracting the peak coordinates of the signal accumulation results based on the hybrid mutual ambiguity function, the coarse time delay error and frequency error to be compensated between radar units can be obtained: (4) Step 3: Based on the search results using the hybrid mutual ambiguity function, after compensating for the frequency error of each radar unit signal, divide the search range for time delay and phase error, and perform signal energy accumulation based on distributed coherent generalized Raytheon-Fourier transform. When the accumulated value is at its maximum, the accurate search results for time delay and phase error are obtained. Step two can address both coarse time delay error and frequency error. The frequency error-compensated signal for each carrier frequency can be expressed as: (5) in, A constant phase that does not change with fast time: (6) Let the time delay error search value be... The phase error search value is Constructing a parameterized coherent synthesized signal model: (7) To improve robustness under low signal-to-noise ratio conditions, signal energy is accumulated in the slow-time dimension based on distributed coherent generalized Raydon-Fourier transform (GFT), and peak energy is extracted in the fast-time dimension. The accumulation expression based on distributed coherent GFT is as follows: (8) When the accumulated value is at its maximum, the search results for accurate time delay and phase error are as follows: (9) Step 4: Based on the searched time delay, frequency, and phase errors, the signals of each radar unit are compensated and then coherently synthesized to obtain an equivalent large-bandwidth signal. Then, the classic delayed Doppler imaging algorithm is applied to obtain high-resolution imaging results. Steps two and three yield accurate results for time delay, frequency, and phase error. Under low signal-to-noise ratio conditions, the amplitude ratio between the signals of each radar element can be represented by the root mean square of the peak energy in the slow time dimension: (10) in, This indicates the peak position after signal accumulation.

[0018] After compensating for amplitude, time delay, frequency, and phase errors, the result of coherent synthesis of the multiple signals is as follows: (11) Example The feasibility and effectiveness of the proposed technology have been verified through computer simulation and experimental data.

[0019] Experiment 1: Computer Simulation The asteroid model used in the simulation is as follows: Figure 3 As shown in the figure. The simulation uses two radar units, each transmitting a 5MHz bandwidth signal with a carrier frequency spacing of 7MHz. The signals of each radar unit are obtained through pulse compression. First, a large-scale, large-step time delay error search grid and a fine-scale frequency error search grid are divided. For each set of error parameters, the accumulation based on the hybrid mutual ambiguity function is calculated. The results are shown in the figure. Figure 4 As shown in Table 1, the error parameter corresponding to the maximum accumulated value is the search result. The search range and results are shown in Table 1. The search results for time delay error and frequency error are close to the true value and remain within one step.

[0020] Table 1 Search results based on mixed mutual ambiguity function

[0021] Secondly, based on the search results from the previous step, after compensating for the frequency error, a small-range, small-step parameter search space is defined near the search results for the coarse time delay error. This space is then combined with the phase error to perform accumulation based on the distributed coherent generalized Rayleigh-Fourier transform. The accumulation results of the echo energy within the parameter space formed by the time delay error and the phase error are shown in the appendix. Figure 5 As shown in Table 2, the time delay error and phase error corresponding to the maximum accumulated value are the search results. The search range and results are shown in Table 2. The search results for time delay error and phase error are basically close to the true values, but since the target's electromagnetic response to different frequencies deviates slightly, some errors still exist in the final search results.

[0022] Table 2 Search results based on distributed coherent generalized Redan-Fourier transform

[0023] Finally, the signals from each radar unit are noncoherently superimposed in a slow time dimension to estimate the signal energy ratio and normalize the amplitude. After compensating for time delay, frequency, and phase errors, the signals from each radar unit are combined into an equivalent large-bandwidth signal. The imaging results after applying the delay-Doppler imaging algorithm are shown in the attached figure. Figure 6 As shown.

[0024] The imaging results show that, without error compensation, the multi-station radar coherent synthesis image exhibits significant range and Doppler direction shifts, directly leading to severe defocusing. After error compensation, the imaging results show good focusing and improved resolution compared to single-station radar imaging.

[0025] Experiment 2: Field Verification of Asteroids The proposed method was validated by conducting field observations of asteroids using the "China Compound Eye" radar system in real-world engineering scenarios.

[0026] Three radar units were used, each transmitting a linear frequency modulated (LFM) signal with a bandwidth of 5MHz, with a carrier frequency spacing of 7MHz. The radar units were numbered 1, 2, and 3, with unit 2 serving as the reference. Using the proposed algorithm, the signals transmitted by radar units 1 and 3 were compared with the signal transmitted by radar unit 2, and the errors were estimated and calibrated. The measured data were segmented, and the error was estimated separately for each segment. The error search results for the entire observation period are shown in the appendix. Figure 7 As shown.

[0027] Based on the error estimation results, the overall trends of time delay and frequency errors of radars 1 and 3 and the reference radar are similar, while no obvious trend was found in the phase error. The imaging results after coherent synthesis, based on the searched time delay, frequency, and phase errors, are shown in the attached figure. Figure 8 As shown, compared with the uncompensated coherent synthesis imaging results, the error-compensated imaging results have better focusing effects. Compared with the monostatic radar imaging results, the target outline in the coherent synthesis imaging results is clearer and more complete. This further verifies the effectiveness of the proposed algorithm.

[0028] In summary, the above are merely preferred embodiments of the present invention and are 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 within the scope of protection of the present invention.

Claims

1. A high-resolution imaging method for deep-space targets based on broadband adaptive coherent synthesis using distributed aperture radar, characterized in that, The method includes the following steps: Step 1: Select the signal transmitted by one radar unit from multiple radar units as the reference signal; after receiving the adjacent carrier frequency narrowband signals transmitted by each radar unit by the same receiving antenna, perform down-conversion processing to baseband and pulse compression to obtain the pulse-compressed signal corresponding to each radar unit. Step 2: Using the reference signal as a reference, construct a joint compensation operator for the fast time of the signal to be calibrated; divide the total observation pulses evenly into N coherent processing intervals, each containing M pulses; use the hybrid mutual ambiguity function HCAF to coherently accumulate the signal to be calibrated and the reference signal within a single coherent processing interval, and perform non-coherent superposition between the N coherent processing intervals; by performing a two-dimensional parameter search on the time delay error and frequency error, obtain the coarse time delay error estimate and frequency error estimate of each radar unit when the accumulated value is at its maximum. Step 3: Perform frequency error compensation on the signals of each radar unit based on the frequency error estimate obtained in Step 2; divide the refined time delay error search range and phase error search range centered on the coarse time delay error estimate obtained in Step 2; use the distributed coherent generalized Radon-Fourier transform (DC-GRFT) to accumulate coherent energy on the signals of each radar unit in the slow time dimension, and extract peak energy in the fast time dimension; by performing a two-dimensional parameter search on the time delay error and phase error, obtain the accurate time delay error estimate and phase error estimate of each radar unit when the accumulated value is the maximum. Step 4: Based on the time delay error, frequency error, and phase error obtained in Steps 2 and 3, perform error compensation and amplitude normalization processing on the pulse compression signals of each radar unit in sequence, and coherently synthesize the signals of each radar unit into an equivalent large bandwidth signal; apply the delayed Doppler imaging algorithm to the equivalent large bandwidth signal to obtain a high-resolution two-dimensional image of the deep space target.

2. The method according to claim 1, characterized in that: Each of the multiple radar units transmits a linear frequency modulated signal, and the carrier frequency interval between the signals transmitted by adjacent radar units is not less than the signal bandwidth of each unit; the reference signal is selected as the signal transmitted by the radar unit with the highest signal-to-noise ratio, or the signal transmitted by a designated radar unit preset by the system; the pulse compression is achieved by matched filtering, and the pulse-compressed signal includes fast time, slow time, signal amplitude, envelope function, frequency modulation slope, frequency error, time delay, carrier frequency, and initial phase parameters.

3. The method according to claim 1, characterized in that, In step two, the pulse compression signal obtained in step one is accumulated based on a hybrid mutual ambiguity function; let the total number of observed pulses be... Divide it evenly into Each coherent processing interval contains _ coherent processing intervals_. One pulse; divide the time delay error search grid into a coarse one and a fine one, and set the time delay error search value as... The frequency offset error search value is Construct a joint compensation operator for the fast time of the path to be calibrated: ； The operator is applied to the signal to be calibrated, performing coherent accumulation within a single coherent processing interval, followed by... Non-coherent superposition is performed between each coherent processing interval; The expression for the hybrid mutual fuzzy function with two-dimensional parameter optimization is as follows: 。 4. The method according to claim 1, characterized in that, In step three, the frequency error-compensated signal of each carrier frequency signal is expressed as: ； in, A constant phase that does not change with fast time: ； Let the time delay error search value be... The phase error search value is Construct a parameterized coherent synthesized signal model: ； The accumulation expression based on the distributed coherent generalized Raydon-Fourier transform is: ； When the accumulated value is at its maximum, the search results for accurate time delay and phase error are as follows: ； The search results for the time delay and phase error corresponding to the maximum accumulated value are as follows: 。 5. The method according to claim 1, characterized in that, In step four, after the error estimates obtained in steps two and three are compensated for, the amplitude ratio of each radar element signal relative to the reference signal under low signal-to-noise ratio conditions is expressed by the root mean square of the peak energy of each element signal in the slow time dimension as follows: ； in, Indicates the peak position after signal accumulation; After compensating for amplitude, time delay, frequency, and phase errors, the result of coherent synthesis of the multiple signals is as follows: 。 6. The method according to claim 5, characterized in that: The coherent synthesis process includes: compressing the pulsed signals of each radar unit to compensate for time delay error, frequency error and phase error, shifting them to the corresponding carrier frequency position, and splicing the signals in the frequency domain into an equivalent large bandwidth signal; the total bandwidth of the equivalent large bandwidth signal after coherent synthesis is equal to the sum of the bandwidths of the signals of each radar unit, and the improvement in imaging range resolution compared to a single radar unit is proportional to the number of radar units participating in the synthesis.

7. The method according to claim 1, characterized in that: The delayed Doppler imaging algorithm performs range-dimensional pulse compression and Doppler-dimensional coherent accumulation processing on the equivalent large bandwidth signal. The horizontal axis of the output two-dimensional image is the time delay dimension, and the corresponding range resolution is determined by the bandwidth of the synthesized equivalent signal; the vertical axis is the Doppler frequency dimension, and the corresponding Doppler resolution is determined by the total coherent accumulation time.

8. A high-resolution imaging system for deep-space targets based on broadband adaptive coherent synthesis distributed aperture radar, characterized in that, The system includes: Multiple radar units are configured to each transmit narrowband linear frequency modulated signals with different carrier frequencies, and the carrier frequency spacing between adjacent radar units is not less than the unit signal bandwidth; The receiving antenna is configured to simultaneously receive echo signals returned after each radar unit illuminates a deep-space target; The signal preprocessing module is configured to perform down-conversion and pulse compression processing on each received echo signal and output the pulse-compressed signal corresponding to each radar unit. The error estimation module includes a first-level estimation submodule and a second-level estimation submodule: the first level uses the hybrid mutual ambiguity function HCAF to estimate the coarse time delay error and frequency error of each radar unit signal; the second level, based on the output of the first level, uses the distributed coherent generalized Radon-Fourier transform DC-GRFT to estimate the precise time delay error and phase error of each radar unit signal. The coherent synthesis module is configured to perform time delay compensation, frequency compensation, phase compensation and amplitude normalization on each signal according to the output of the error estimation module, and to combine each signal into an equivalent large bandwidth signal in the frequency domain. The imaging processing module is configured to perform a delayed Doppler imaging algorithm on the equivalent large bandwidth signal and output a high-resolution two-dimensional time-delay-Doppler image of the deep space target.

Images