Radar echo map prediction method and system, terminal and storage medium
By combining mechanism-based signal-level radar model and data-driven model, the problem of the inability to accurately predict radar echo maps in the existing technology is solved, and the accurate and real-time prediction of radar echo maps is achieved.
Patent Information
- Application Number
- CN202411986821.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-31
- Publication Date
- 2025-06-03
AI Technical Summary
The mechanism-based models in the prior art cannot accurately predict all features in the radar echo map, resulting in the inability to accurately and in real time to predict the echo map data output by the radar.
A signal-level radar model based on mechanism architecture is used as a prior model, and a prior echo graph is output based on scene information, and input it into data-driven model (such as Unet network) for prediction, and output the real echo graph of the radar.
The main features of the echo graph are generated through the prior model and the detailed parts are adjusted by the data-driven model, which significantly improves the prediction accuracy of the echo graph and achieves accurate and real-time radar echo graph prediction.
Smart Images
Figure CN120087174A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and particularly to a method, system, terminal and computer-readable storage medium for predicting radar echo images. Background Art
[0002] A radar echo image contains a large amount of complex information. Predicting the echo image based on scene information by a radar is a complex task. However, the existing mechanism-based models cannot accurately predict all the features in the echo image, and thus cannot accurately and real-time predict the echo image data output by the radar.
[0003] Therefore, the existing technology still needs to be improved and developed. Summary of the Invention
[0004] The main purpose of the present invention is to provide a method, system, terminal and computer-readable storage medium for predicting radar echo images, aiming at solving the problem that in the prior art, when predicting an echo image, the mechanism-based model cannot accurately predict all the features in the echo image, and thus cannot accurately and real-time predict the echo image data output by the radar.
[0005] To achieve the above object, the present invention provides a method for predicting a radar echo image, the method for predicting the radar echo image comprising the following steps:
[0006] Obtain scene information, and use a signal-level radar model based on a mechanism architecture as a prior model, the prior model outputting a prior echo image according to the scene information;
[0007] Input the prior echo image into a data-driven model, and use the data-driven model to predict the prior echo image to output a true echo image of the radar.
[0008] Optionally, in the method for predicting a radar echo image, the signal-level radar model based on a mechanism architecture includes an ideal radio frequency model and a scattering center model.
[0009] Optionally, in the method for predicting a radar echo image, the step of obtaining scene information, using a signal-level radar model based on a mechanism architecture as a prior model, and the prior model outputting a prior echo image according to the scene information specifically includes:
[0010] Obtain scene information, obtain scene ground truth according to the scene information, and input the scene ground truth into the scattering center model;
[0011] The scattering center model calculates a radar echo signal according to the scene ground truth, and sends the echo signal to the ideal radio frequency model;
[0012] The ideal radio frequency model processes the received echo signal and outputs a prior echo map.
[0013] Optionally, in the method for predicting a radar echo map, the step of inputting the prior echo map into a data-driven model and using the data-driven model to predict the prior echo map to output a true echo map of the radar specifically includes:
[0014] The data-driven model receives the prior echo map sent by the ideal radio frequency model, and the data-driven model uses a Unet network to predict the prior echo map and outputs a true echo map of the radar.
[0015] Optionally, in the method for predicting a radar echo map, the method for predicting a radar echo map further includes:
[0016] If the radar only receives the signal reflected by the target equivalent point, the frequency of the signal is obtained from the frequency of the radar transmitted signal and the radial velocity of the target equivalent point, the relative phase value between different receiving channels is calculated according to the element spacing, and the delay of the signal returning to the radar at the target equivalent point is calculated according to the distance of the target:
[0017]
[0018] ΔT = R length / c; (3)
[0019] where f echo represents the frequency of the echo signal, f tx represents the frequency of the radar transmitted signal, v r represents the radial relative velocity between the radar and the target, λ represents the wavelength of the electromagnetic wave, represents the phase difference between different receiving channels, d represents the spacing between different receiving elements, θ represents the relative angle between the target and the radar, ΔT represents the delay of the signal returning to the radar at the target equivalent point, R length represents the total distance between the target and the radar, and c represents the speed of light.
[0020] Optionally, in the method for predicting a radar echo map, the ideal radio frequency model includes a signal source, a power splitter, a transmitting antenna, a receiving antenna, a mixer, and an AD conversion module;
[0021] The signal source is used to set the parameters of the transmitted signal;
[0022] The power splitter is used to divide the signal into two parts, one part is transmitted to the transmitting antenna, and the other part is transmitted to the mixer as the signal of this frame;
[0023] The transmitting antenna is used to transmit the signal;
[0024] The receiving antenna is used to receive signals;
[0025] The mixer is used to subtract the frequency of the current frame signal from the radio frequency signal received by the receiving antenna to obtain an intermediate frequency signal, and sample it using the AD conversion module to obtain raw data.
[0026] Optionally, in the radar echo map prediction method, the prior echo map includes a range-angle echo map and a range-Doppler echo map;
[0027] The ideal radio frequency model processes the received echo signal and outputs a prior echo map, specifically including:
[0028] Perform fast Fourier transform calculation on the fast time dimension of the raw data to extract the range information in the data;
[0029] Perform fast Fourier transform on the data output after fast Fourier transform in the slow time dimension to extract the velocity information;
[0030] Perform non-coherent accumulation on the data that has undergone two-dimensional fast Fourier transform to obtain a range-Doppler echo map;
[0031] Perform fast Fourier transform on the data after the slow time dimension in the spatial dimension to extract the angle information, and downsample the three-dimensional data matrix of the range information, velocity information, and angle information to obtain a range-angle echo map;
[0032] Wherein, the fast time dimension refers to the echo signal of a single pulse emission period, which contains the range information of the target; the slow time dimension refers to the time interval between consecutive pulse emissions, which contains the velocity information of the target.
[0033] In addition, to achieve the above object, the present invention also provides a radar echo map prediction system, wherein the radar echo map prediction system includes:
[0034] A prior echo map processing module, used to obtain scene information, and use the signal-level radar model based on the mechanism architecture as the prior model, and the prior model outputs a prior echo map according to the scene information;
[0035] A real echo map prediction module, used to input the prior echo map into the data-driven model, and use the data-driven model to predict the prior echo map and output the real echo map of the radar.
[0036] In addition, to achieve the above object, the present invention further provides a terminal, wherein the terminal includes: a memory, a processor, and a prediction program of a radar echo map stored on the memory and executable on the processor. When the prediction program of the radar echo map is executed by the processor, the steps of the above-described prediction method of the radar echo map are implemented.
[0037] In addition, to achieve the above object, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a prediction program of a radar echo map. When the prediction program of the radar echo map is executed by a processor, the steps of the above-described prediction method of the radar echo map are implemented.
[0038] In the present invention, scene information is obtained, and a signal-level radar model based on a mechanism architecture is used as a prior model. The prior model outputs a prior echo map according to the scene information; the prior echo map is input into a data-driven model, and the data-driven model is used to predict the prior echo map to output a true echo map of the radar. The present invention uses a signal-level radar model based on a mechanism architecture as a prior model, takes advantage of its advantages in real-time performance, high generalization ability, and prediction accuracy of main features, generates the backbone features in the echo map according to the scene information, and the features of the prior model ensure the accuracy of the backbone features of the echo map, avoiding deviation from the true value, and providing a reliable prediction basis for the data-driven model. The data-driven model is used to quickly adjust the details of the echo map. Relying on its high accuracy and robustness in simple mapping relationship problems, the data-driven model takes the echo map generated by the prior model as input and predicts a result closer to the true echo map, so as to accurately and real-time predict the echo map data output by the radar. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 is a flowchart of a preferred embodiment of the prediction method of the radar echo map of the present invention;
[0040] Figure 2 is a schematic diagram of a signal-level radar model based on a mechanism architecture in a preferred embodiment of the prediction method of the radar echo map of the present invention;
[0041] Figure 3 is a schematic diagram of the prior model and the data-driven model completing the prediction of the radar echo map in a preferred embodiment of the prediction method of the radar echo map of the present invention;
[0042] Figure 4 is a comparison schematic diagram of the output of the Unet model and the echo map of the ideal mechanism model in a preferred embodiment of the prediction method of the radar echo map of the present invention;
[0043] Figure 5 is a structural diagram of a preferred embodiment of the prediction system of the radar echo map of the present invention;
[0044] Figure 6 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed implementation manners
[0045] To make the objectives, technical solutions and advantages of the present invention clearer and more explicit, the following further describes the present invention in detail with reference to the accompanying drawings and by way of examples. It should be understood that the specific examples described herein are only used to explain the present invention and are not used to limit the present invention.
[0046] The method for predicting a radar echo map according to a preferred embodiment of the present invention, as Figure 1 and Figure 3 shown, the method for predicting a radar echo map includes the following steps:
[0047] Step S10: Obtain scene information, use a signal-level radar model based on a mechanism architecture as a prior model, and the prior model outputs a prior echo map according to the scene information.
[0048] Specifically, as Figure 2 shown, the signal-level radar model based on a mechanism architecture includes an ideal radio frequency model and a scattering center model; obtain scene information, obtain a scene ground truth according to the scene information, and input the scene ground truth into the scattering center model; the scattering center model calculates the echo signal of the radar according to the scene ground truth and sends the echo signal to the ideal radio frequency model; the ideal radio frequency model processes the received echo signal and outputs a prior echo map.
[0049] The method of the present invention can be used in tasks such as radar modeling, and further can be used in simulation systems for autonomous driving and unmanned aerial vehicles. For example, when applied to autonomous driving, a signal-level radar model is established based on a mechanism architecture. For example, using the scattering center method, the target vehicle is equivalent to a point target at the center position of the vehicle. This method ignores the multipath effect of radar signals and excludes the influence of other factors in the scene (such as ground reflection and fixed objects), assuming that the radar only receives the signal reflected by the target equivalent point. The frequency of the signal is obtained by the frequency of the radar transmitted signal and the radial velocity of the target equivalent point. This method usually does not calculate the absolute value of the phase of the target, but only calculates the relative phase value between different receiving channels according to the element spacing. In addition, according to the distance of the target, the delay of the signal returned to the radar at the target equivalent point is calculated:
[0050]
[0051] ΔT = R length / c; (3)
[0052] where, f echo represents the frequency of the echo signal, f txdenotes the frequency of the radar transmitted signal, v r denotes the radial relative velocity between the radar and the target, λ denotes the wavelength of the electromagnetic wave, denotes the phase difference between different receiving channels, d denotes the spacing between different receiving array elements, θ denotes the relative angle between the target and the radar, ΔT denotes the delay of the signal returned to the radar at the equivalent point of the target, R length denotes the total distance between the target and the radar, c denotes the speed of light.
[0053] The parameters of the signal received by the radar are calculated using the above formulas (1)-(3). It is Figure 2 the output of the scattering center model.
[0054] As Figure 2 shown, the scattering center model is used to generate the echo signal received by the radar for the test scenario; the ideal radio frequency model includes a signal source, a power splitter, a transmitting antenna (TX), a receiving antenna (RX), a mixer, and an AD conversion module (ADC); the signal source is used to set parameters such as the frequency of the transmitted signal; the power splitter is used to divide the signal into two parts, one part is transmitted to the transmitting antenna, and the other part is transmitted to the mixer as the local oscillator signal (LO) of this frame; the transmitting antenna is used to transmit the signal; the receiving antenna is used to receive the signal; the mixer is used to subtract the frequency of the local oscillator signal (LO) from the frequency of the radio frequency signal (RF) received by the receiving antenna to obtain an intermediate frequency signal, and sample it using the AD conversion module to obtain the raw data Raw Data.
[0055] After performing FFT transform (Fast Fourier Transform) on the raw data Raw Data, the prior echo map of the radar is obtained, and the prior echo map includes a range-angle map and a range-doppler map.
[0056] Perform fast Fourier transform (FFT transform) calculation on the fast time dimension of the raw data to extract the range information in the data; then perform fast Fourier transform (FFT transform) on the data output after performing fast Fourier transform (FFT transform) on the slow time dimension to extract the velocity information; perform non-coherent accumulation on the data that has undergone fast Fourier transform (FFT transform) in two dimensions to obtain a range-doppler map; finally, perform fast Fourier transform (FFT transform) on the data after the slow time dimension in the spatial dimension to extract the angle information, and downsample the three-dimensional data matrix of the range information, velocity information, and angle information to obtain a range-angle map.
[0057] Among them, the fast time dimension refers to the echo signal of a single pulse transmission period, which contains the distance information of the target; the slow time dimension refers to the time interval between consecutive pulse transmissions, which contains the velocity information of the target.
[0058] Step S20: Input the prior echo map into the data-driven model, and use the data-driven model to predict the prior echo map to output the true echo map of the radar.
[0059] Specifically, the data-driven model receives the prior echo map sent by the ideal radio frequency model, and the data-driven model uses the Unet network to predict the prior echo map to output the true echo map of the radar.
[0060] As Figure 3 shown, the prior echo map output by the prior model is used as the input, and the data-driven model is used to fit the true echo map. The present invention uses the Unet network to fit the relationship between the prior echo map and the true echo map, making up for the deficiency of the prior model in predicting detailed features.
[0061] The present invention uses the signal-level radar model based on the mechanism architecture as the prior model, takes advantage of its advantages in real-time performance, high generalization ability, and prediction accuracy of main features, generates the backbone features in the echo map according to the scene information, and the features of the prior model ensure the accuracy of the backbone features of the echo map, avoiding deviation from the true value and providing a reliable prediction basis for the data-driven model. Subsequently, the data-driven model quickly adjusts the detailed part of the prior echo map. Relying on its high accuracy and robustness in simple mapping relationship problems, the data-driven model uses the prior echo map generated by the prior model as the input to predict a result closer to the true echo map. In this way, the data-driven model makes up for the deficiency of the prior model in predicting detailed features, thereby significantly improving the prediction accuracy of the echo map.
[0062] The present invention can effectively improve the prediction accuracy of the model. Conventional mechanism-based methods are limited by their own limitations and cannot accurately calculate the radar echo map. However, the present invention uses a data-driven method to improve the echo map output by the mechanism model, enhancing the accuracy of the echo map. At the same time, it avoids the end-to-end approach of directly using the scene true value to predict the radar echo map in conventional data-driven methods, which can significantly improve the confidence of the radar model in different simulations and promote the development of simulation tests such as autonomous driving and unmanned aerial vehicles.
[0063] Next, the implementation of the prediction method for the radar echo map is further introduced:
[0064] The first step is to establish a radar model based on the mechanism (i.e., the signal-level radar model based on the mechanism architecture) for outputting prior echo map data.
[0065] Calculate the amplitude, frequency, and time delay of the target echo signal according to the formula. Each component in the signal-level radar model based on the mechanism architecture is an ideal component. Set the operating frequency to 77 GHz, the bandwidth to 400 MHz, and the period to 20 us. The transmitting antenna and the receiving antenna usually use non-directional omnidirectional antennas. The mixer mixes the local oscillator signal and the RF signal according to the delay of the received signal to calculate the amplitude, frequency, and phase information of the intermediate frequency signal. Then, the calculated signal is processed by FFT to obtain the echo map.
[0066] In the second step, use the prior echo map to train the Unet network (the Unet network is used in the data-driven model).
[0067] Use the method proposed in the present invention to establish a radar model. The training hyperparameters of the present invention are shown in the following table:
[0068]
[0069]
[0070] The echo map obtained thereby is as Figure 4 shown. It can be seen that the output of the Unet model has a significant improvement in the prediction of detailed features compared to the echo map of the ideal mechanism model. The predicted echo map has improved in the target distribution area. The MSE (Mean Squared Error) of the range-angle map (i.e., the Range-Angle Map in Figure 4 ) and the range-doppler map (Range-Doppler Map) has decreased significantly compared to the ideal mechanism model.
[0071] Furthermore, as Figure 5 shown, based on the above-mentioned radar echo map prediction method, the present invention also correspondingly provides a radar echo map prediction system, wherein the radar echo map prediction system includes:
[0072] A prior echo map processing module 51, configured to obtain scene information, use the signal-level radar model based on the mechanism architecture as a prior model, and the prior model outputs a prior echo map according to the scene information;
[0073] A real echo map prediction module 52, configured to input the prior echo map into the data-driven model, use the data-driven model to predict the prior echo map, and output the real echo map of the radar.
[0074] Furthermore, as Figure 6As shown, based on the above radar echo map prediction method and system, the present invention also correspondingly provides a terminal, which includes a processor 10, a memory 20, and a display 30. Figure 6 Only some components of the terminal are shown, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be alternatively implemented.
[0075] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as the hard disk or memory of the terminal. In some other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard disk equipped on the terminal, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. Further, the memory 20 may also include both the internal storage unit and the external storage device of the terminal. The memory 20 is used to store application software installed on the terminal and various types of data, such as the program code installed on the terminal, etc. The memory 20 may also be used to temporarily store data that has been output or will be output. In one embodiment, a prediction program 40 of the radar echo map is stored on the memory 20, and the prediction program 40 of the radar echo map can be executed by the processor 10, thereby implementing the radar echo map prediction method in the present application.
[0076] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chips, and is used to run the program code stored in the memory 20 or process data, such as executing the radar echo map prediction method, etc.
[0077] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch liquid crystal display, and an OLED (Organic Light-Emitting Diode) toucher, etc. The display 30 is used to display information on the terminal and is used to display a visual user interface. The processor 10, the memory 20, and the display 30 of the terminal communicate with each other through a system bus.
[0078] In one embodiment, when the processor 10 executes the prediction program 40 of the radar echo map in the memory 20, the following steps are implemented:
[0079] Obtain scene information, use the signal-level radar model based on the mechanism architecture as a prior model, and the prior model outputs a prior echo map according to the scene information;
[0080] Input the prior echo map into a data-driven model, and use the data-driven model to predict the prior echo map to output the true echo map of the radar.
[0081] Among them, the mechanism architecture-based signal-level radar model includes an ideal radio frequency model and a scattering center model.
[0082] Among them, for obtaining scene information and using the mechanism architecture-based signal-level radar model as a prior model, the prior model outputs a prior echo map according to the scene information, specifically including:
[0083] Obtain scene information, obtain the scene ground truth according to the scene information, and input the scene ground truth into the scattering center model;
[0084] The scattering center model calculates the echo signal of the radar according to the scene ground truth and sends the echo signal to the ideal radio frequency model;
[0085] The ideal radio frequency model processes the received echo signal and outputs a prior echo map.
[0086] Among them, for inputting the prior echo map into a data-driven model and using the data-driven model to predict the prior echo map to output the true echo map of the radar, specifically:
[0087] The data-driven model receives the prior echo map sent by the ideal radio frequency model, and the data-driven model uses the Unet network to predict the prior echo map to output the true echo map of the radar.
[0088] Among them, the prediction method of the radar echo map further includes:
[0089] If the radar only receives the signal reflected by the target equivalent point, the frequency of the signal is obtained from the frequency of the radar transmitted signal and the radial velocity of the target equivalent point, calculate the relative phase value between different receiving channels according to the element spacing, and calculate the delay of the signal returning to the radar at the target equivalent point according to the distance of the target:
[0090]
[0091] ΔT = R length / c; (3)
[0092] Among them, f echo represents the frequency of the echo signal, f tx represents the frequency of the radar transmitted signal, v r represents the radial relative velocity between the radar and the target, and λ represents the wavelength of the electromagnetic wave, It represents the phase difference between different receiving channels, d represents the spacing between different receiving array elements, θ represents the relative angle between the target and the radar, ΔT represents the delay of the signal returned to the radar at the equivalent point of the target, and R length represents the total distance between the target and the radar, and c represents the speed of light.
[0093] Among them, the ideal RF model includes a signal source, a power splitter, a transmitting antenna, a receiving antenna, a mixer, and an AD conversion module;
[0094] The signal source is used to set the parameters of the transmitted signal;
[0095] The power splitter is used to divide the signal into two parts, one part is transmitted to the transmitting antenna, and the other part is transmitted to the mixer as the signal of this frame;
[0096] The transmitting antenna is used to transmit the signal;
[0097] The receiving antenna is used to receive the signal;
[0098] The mixer is used to subtract the frequency of the signal of this frame from the frequency of the RF signal received by the receiving antenna to obtain an intermediate frequency signal, and use the AD conversion module for sampling to obtain the original data.
[0099] Among them, the prior echo map includes a range-angle echo map and a range-Doppler echo map;
[0100] The ideal RF model processes the received echo signal and outputs a prior echo map, specifically including:
[0101] Perform fast Fourier transform calculation on the fast time dimension of the original data to extract the range information in the data;
[0102] Perform fast Fourier transform on the data output after fast Fourier transform in the slow time dimension to extract the velocity information;
[0103] Perform non-coherent accumulation on the data that has undergone fast Fourier transform in two dimensions to obtain a range-Doppler echo map;
[0104] Perform fast Fourier transform on the data in the slow time dimension in the spatial dimension to extract the angle information, and downsample the three-dimensional data matrix of the range information, velocity information, and angle information to obtain a range-angle echo map;
[0105] Among them, the fast time dimension refers to the echo signal of a single pulse transmission period, which contains the range information of the target; the slow time dimension refers to the time interval between consecutive pulse transmissions, which contains the velocity information of the target.
[0106] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a prediction program for radar echo maps, and when the prediction program for radar echo maps is executed by a processor, the steps of the prediction method for radar echo maps as described above are implemented.
[0107] In summary, the present invention provides a method, a system, a terminal, and a computer-readable storage medium for predicting radar echo maps. The method includes: obtaining scene information, using a signal-level radar model based on a mechanism architecture as a prior model, and the prior model outputs a prior echo map according to the scene information; inputting the prior echo map into a data-driven model, and using the data-driven model to predict the prior echo map to output a true echo map of the radar. The present invention uses a signal-level radar model based on a mechanism architecture as a prior model, takes advantage of its superiority in real-time performance, high generalization ability, and prediction accuracy of main features, generates the backbone features in the echo map according to the scene information, and the features of the prior model ensure the accuracy of the backbone features of the echo map, avoiding deviation from the true value, and providing a reliable prediction basis for the data-driven model. The data-driven model is used to quickly adjust the details of the echo map. Relying on its high accuracy and robustness in simple mapping relationship problems, the data-driven model takes the echo map generated by the prior model as input and predicts a result closer to the true echo map, so as to accurately and real-time predict the echo map data output by the radar.
[0108] It should be noted that in this article, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article or terminal. Without more limitations, the element defined by the statement "including one..." does not exclude the existence of another identical element in the process, method, article or terminal including that element.
[0109] Certainly, those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware (such as a processor, a controller, etc.) through a computer program. The program can be stored in a computer-readable storage medium that can be read by a computer. When the program is executed, it can include the processes of the above method embodiments. The computer-readable storage medium can be a memory, a magnetic disk, an optical disk, etc.
[0110] It should be understood that the application of the present invention is not limited to the above examples. For those of ordinary skill in the art, improvements or transformations can be made according to the above description, and all such improvements and transformations should fall within the protection scope of the appended claims of the present invention.
Claims
1. A method for predicting a radar echogram, characterized in that: The radar echogram prediction method comprises: Acquire scene information, and use a signal-level radar model based on a mechanism architecture as a priori model, wherein the priori model outputs a priori echo map according to the scene information; The priori echo map is input into a data-driven model, the data-driven model is used to predict the priori echo map, and a real echo map of the radar is output.
2. The radar echogram prediction method according to claim 1, characterized in that: The signal-level radar model based on the mechanism architecture includes an ideal radio frequency model and a scattering center model.
3. The radar echogram prediction method according to claim 2, characterized in that: The acquiring of scene information, using a signal-level radar model based on a mechanism architecture as a priori model, wherein the priori model outputs a priori echo map according to the scene information, specifically includes: Acquire scene information, acquire scene truth value according to the scene information, and input the scene truth value into the scattering center model; The scattering center model calculates the radar echo signal according to the scene true value, and sends the echo signal to the ideal radio frequency model; The ideal radio frequency model processes the received echo signal and outputs a priori echo map.
4. The radar echogram prediction method according to claim 3, characterized in that: The priori echo map is input into the data driven model, the priori echo map is predicted using the data driven model, and the real echo map of the radar is output, specifically: The data-driven model receives the priori echo map sent by the ideal radio frequency model, and the data-driven model predicts the priori echo map using a Unet network to output a real echo map of the radar.
5. The radar echogram prediction method according to claim 1, characterized in that: The radar echogram prediction method further includes: If the radar only receives the signal reflected from the target equivalent point, the frequency of the signal is obtained by the frequency of the radar transmitting signal and the radial velocity of the target equivalent point. The relative phase values between different receiving channels are calculated according to the array element spacing, and the delay of the signal returned to the radar at the target equivalent point is calculated according to the target distance: ΔT=R length / c; (3) Among them, f echo Indicates the frequency of the echo signal, f tx represents the frequency of the radar transmitting signal, v r represents the radial relative speed between the radar and the target, λ represents the wavelength of the electromagnetic wave, represents the phase difference between different receiving channels, d represents the spacing between different receiving array elements, θ represents the relative angle between the target and the radar, ΔT represents the delay of the signal returning to the radar at the target equivalent point, and R length represents the total distance between the target and the radar, and c represents the speed of light.
6. The radar echogram prediction method according to claim 3, characterized in that: The ideal radio frequency model includes a signal source, a power divider, a transmitting antenna, a receiving antenna, a mixer and an AD conversion module; The signal source is used to set the parameters of the transmission signal; The power divider is used to divide the signal into two parts, one part is transmitted to the transmitting antenna, and the other part is transmitted to the mixer as the current frame signal; The transmitting antenna is used to transmit signals; The receiving antenna is used to receive signals; The mixer is used to subtract the frequency of the current frame signal from the frequency of the radio frequency signal received by the receiving antenna to obtain an intermediate frequency signal, and use the AD conversion module to perform sampling to obtain original data.
7. The radar echogram prediction method according to claim 6, characterized in that: The priori echo map includes a distance-angle echo map and a distance-Doppler echo map; The ideal radio frequency model processes the received echo signal and outputs a priori echo map, specifically including: Performing a fast Fourier transform calculation on the fast time dimension of the original data to extract distance information from the data; Perform fast Fourier transform on the data output by fast Fourier transform in the slow time dimension to extract the speed information; The data subjected to two-dimensional fast Fourier transform are incoherently accumulated to obtain a range-Doppler echogram; Performing fast Fourier transform on the data after the slow time dimension in the spatial dimension to extract the angle information, and down-sampling the three-dimensional data matrix of the distance information, speed information and angle information to obtain the distance-angle echo map; The fast time dimension refers to the echo signal of a single pulse transmission cycle, which contains the distance information of the target; the slow time dimension refers to the time interval between consecutive pulse transmissions, which contains the speed information of the target.
8. A radar echo pattern prediction system, characterized in that: The radar echogram prediction system comprises: A priori echo map processing module is used to obtain scene information, and use a signal-level radar model based on a mechanism architecture as a priori model, wherein the priori model outputs a priori echo map according to the scene information; The real echo map prediction module is used to input the priori echo map into the data-driven model, use the data-driven model to predict the priori echo map, and output the real echo map of the radar.
9. A terminal, characterized in that: The terminal includes: a memory, a processor, and a radar echo map prediction program stored in the memory and executable on the processor. When the radar echo map prediction program is executed by the processor, the steps of the radar echo map prediction method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a radar echo map prediction program, and when the radar echo map prediction program is executed by a processor, the steps of the radar echo map prediction method according to any one of claims 1 to 7 are implemented.