A method for suppressing sidelobes of a single-bit quantized chirp signal by pulse compression

Abstract

The application discloses a method for sidelobe suppression of single-bit quantization Chirp signal, and the main technical content of the method comprises sidelobe suppression of single-bit quantization Chirp signal based on a convolution window function. The technical content performs sidelobe suppression through the following steps: step S1, generating a Chirp echo signal; step S2, single-bit quantization of the Chirp signal; step S3, single-bit quantization of a Chirp signal matching filter and adding a convolution window; and step S4, pulse compression of the single-bit quantization Chirp signal. The application can achieve a good compromise between quantization order and pulse compression effect, single-bit quantization can reduce quantization complexity to the maximum extent, and the sidelobe suppression method based on the convolution window function can effectively solve the problem of the decrease of the main sidelobe ratio caused by single-bit quantization. For Chirp signals and other radar signals with large time-bandwidth product, the application can improve the pulse compression effect, improve the range resolution of the radar system, and has certain reference significance for improved design in radar engineering technology.

Description

Technical Field

[0001] This invention relates to a method for suppressing pulse compression sidelobes in single-bit quantized Chirp signals, belonging to the field of space communication and navigation technology. Background Technology

[0002] Linear frequency modulated (Chirp) signals have a large time-bandwidth product, are easy to generate, and are technically mature. Furthermore, matched filters are insensitive to the Doppler shift of their echo signals, making Chirp the earliest and most widely used pulse compression signal studied in radar systems. Pulse compression technology can effectively resolve the conflict between detection capability and range resolution, and possesses potential anti-jamming capabilities, leading to its widespread application in modern radar systems.

[0003] Currently, it is easy to obtain chirp signals with a large bandwidth (hundreds of megabits) at the transmitting end. However, at the receiving end, due to the limitations of AD devices, the sampling rate often cannot satisfy the Nyquist sampling theorem. Furthermore, due to the cumbersome Discrete Fourier Transform (DFT) operation, traditional channelized receivers have relatively slow processing speeds, creating a certain contradiction between high-speed sampling and real-time signal processing. Currently, single-bit quantization technology is widely used in engineering, which avoids multiplication operations in the DFT, effectively improving processing speed, reducing system complexity, and effectively alleviating the contradiction between high-speed sampling and real-time signal processing. This also improves the overall size and resource consumption of ultra-wideband signal receivers to some extent.

[0004] Chirp signals are compressed into narrow pulses after matched filtering to achieve maximum signal-to-noise ratio. However, compression inevitably generates gradually decreasing sidelobes with a sinc function envelope on both sides of the narrow pulse. While single-bit quantization technology offers advantages in receiver processing, it increases pulse compression sidelobes and reduces the main-to-side-lobe ratio. In multi-target environments, the main signal of smaller targets can be overwhelmed by strong sidelobes, leading to target loss. To improve multi-target resolution, single-bit quantization can significantly reduce the computational load of DFT and lower receiver power consumption and complexity. However, it significantly worsens the pulse compression main-to-side-lobe ratio of linear frequency modulated signals, easily causing false alarms or missed detections of weak targets near the sidelobes. Summary of the Invention

[0005] The technical problem solved by this invention is to overcome the shortcomings of the prior art and provide a pulse compression sidelobe suppression method for single-bit quantized Chirp signals. To address this problem, this invention studies how to suppress the pulse compression sidelobe peak of linear frequency modulated signals after single-bit quantization, improve the main-to-sidelobe ratio, and thereby reduce the probability of missed or false alarms for weak targets near the sidelobes.

[0006] The technical solution of this invention is:

[0007] This invention discloses a method for suppressing pulse compression sidelobes in single-bit quantized chirp signals, comprising:

[0008] Step S1: Generate Chirp echo signal;

[0009] Step S2: Perform single-bit quantization on the Chirp echo signal to obtain the quantized Chirp echo signal;

[0010] Step S3: Perform single-bit quantization on the conjugate matched filter signal of the Chirp echo signal and add a convolution window to obtain a new matched filter signal;

[0011] Step S4: Use a new matched filter signal to perform pulse compression on the quantized Chirp echo signal to obtain the received signal.

[0012] Furthermore, in the above suppression method, the step of performing single-bit quantization on the Chirp echo signal to obtain the quantized signal specifically involves: performing single-bit encoding on the I and Q channels of the Chirp echo signal using a comparator; if the original value of the I or Q channel signal is positive, it is quantized to the value 1; if the original value is negative, it is quantized to the value -1.

[0013] Furthermore, in the above suppression method, the step of performing single-bit quantization on the conjugate matched filter signal of the Chirp echo signal and adding a convolution window to obtain a new matched filter signal specifically involves:

[0014] The conjugate matched filter signal of the original echo signal is quantized by a single bit to obtain the quantized conjugate matched filter signal;

[0015] Based on the frequency domain characteristics of various window functions and the requirements of actual application scenarios, simulation is used to determine the type and number of window functions participating in convolution, thus obtaining the selected convolution window functions.

[0016] The obtained convolutional window function is used to window and truncate the quantized conjugate matched filter signal to obtain a new matched filter signal.

[0017] Furthermore, in the above suppression method, the step of using a new matched filter signal to pulse compress the quantized Chirp echo signal to obtain the received signal specifically involves multiplying the new matched filter signal with the quantized Chirp echo signal to obtain the pulse-compressed received signal.

[0018] Furthermore, in the above suppression method, the step of determining the type and number of window functions participating in convolution based on the frequency domain characteristics of various window functions and the needs of actual application scenarios through simulation, thereby obtaining the selected convolution window function, specifically involves:

[0019] S1. Calculate the spectral function of the cosine window;

[0020] S2. Calculate the frequency domain characteristics of the first-order single window function based on the spectrum function, and determine whether the frequency domain characteristics meet the requirements. If not, proceed to step S3; if yes, obtain the selected convolution window function.

[0021] S3. Calculate the frequency domain characteristics of the Nth-order single window function and determine whether the frequency domain characteristics meet the requirements. If not, proceed to step S4; if yes, obtain the selected convolution window function.

[0022] S4. Calculate the frequency domain characteristics of the mixed convolution of rectangular window with triangular window, Hanning window, Hamming window or Blackman window; determine whether the frequency domain characteristics meet the requirements. If not, proceed to step S5; if yes, obtain the selected convolution window function.

[0023] S5. Calculate the frequency domain characteristics of the mixed convolution of a first-order rectangular window with an Nth-order triangular window, Hanning window, Hamming window, or Blackman window; select the spectral characteristics that best meet the requirements, and the corresponding convolution window is the selected convolution window function; where N is an integer less than or equal to 4.

[0024] Furthermore, in the above suppression method, the specific method for calculating the spectral function of the cosine window is as follows:

[0025]

[0026]

[0027] In the formula, L is the number of cosine terms, and a l is the coefficient of the cosine combination term, and M is the length of the cosine window.

[0028] Furthermore, in the above suppression method, the frequency domain characteristics of the first-order single window function are specifically as follows:

[0029] Rectangular window: Window length is N, main lobe width is 4π / N, and side lobe peak value is less than -13.31dB;

[0030] Triangular window: Window length is N, main lobe width is 8π / N, and side lobe peak value is less than -26.63dB;

[0031] Hanning window: Window length is N, main lobe width is 8π / N, and side lobe peak value is less than -31.5dB;

[0032] Hamming window: Window length is N, main lobe width is 8π / N, and side lobe peak value is less than -44.7dB;

[0033] Blackman window: window length is N, main lobe width is 12π / N, and side lobe peak value is less than -58.22dB.

[0034] Furthermore, in the above suppression method, the frequency domain characteristics of the Nth-order single window function are specifically as follows:

[0035] Rectangular window: When the order is 2 to 4, the window length is N, the main lobe width is 4π / N, and the side lobe peak value is -53.25 to -26.62 dB;

[0036] For triangular windows of order 2 to 4, the window length is N, the main lobe width is 8π / N, and the side lobe peak value is -106.6 to -53.25 dB.

[0037] Hanning window: When the order is 2 to 4, the window length is N, the main lobe width is 8π / N, and the side lobe peak value is -126 to -63 dB;

[0038] Hamming window: When the order is 2 to 4, the window length is N, the main lobe width is 8π / N, and the side lobe peak value is -178.8 to -89.41 dB;

[0039] Blackman window: When the order is 2 to 4, the window length is N, the main lobe width is 12π / N, and the side lobe peak value is -232.9 to -174.7 dB.

[0040] Furthermore, in the above suppression method, the frequency domain characteristics of the mixed convolution of the rectangular window with the triangular window, Hanning window, Hamming window, or Blackman window are specifically as follows:

[0041] Using a hybrid convolution with rectangular and triangular windows, the sidelobe peak value is less than -28.34dB, and the main lobe width is 4π / N;

[0042] Using a hybrid convolution of rectangular and Hanning windows, the sidelobe peak value is less than -24.96dB, and the main lobe width is 4π / N;

[0043] Using a hybrid convolution of rectangular and Hamming windows, the sidelobe peak value is less than -27.54dB, and the main lobe width is 4π / N;

[0044] Using a hybrid convolution of rectangular and Blackman windows, the sidelobe peak value is less than -21.83dB, and the main lobe width is 4π / N.

[0045] Furthermore, in the above suppression method, the frequency domain characteristics of the mixed convolution of the first-order rectangular window with the Nth-order triangular window, Hanning window, Hamming window, or Blackman window are specifically as follows:

[0046] The hybrid convolution pulse compression results of first-order rectangular window and Nth-order triangular window suppress sidelobes to -28.34dB to -62.06dB;

[0047] The hybrid convolution pulse compression results of the first-order rectangular window and the Nth-order Hanning window suppress the sidelobes to -24.96dB to -51.01dB;

[0048] The hybrid convolution pulse compression results of first-order rectangular window and N-order Hamming window suppress sidelobes to -27.54dB to -58.92dB;

[0049] The hybrid convolution pulse compression results of first-order rectangular window and N-order Blackman window suppress sidelobes to -21.83dB to -42dB;

[0050] Where N is an integer less than or equal to 4.

[0051] The advantages of this invention over the prior art are as follows:

[0052] (1) After the chirp signal is quantized, the introduction of quantization error will affect the pulse compression output, leading to a decrease in the main lobe-to-side lobe ratio. To obtain a good pulse compression effect, the quantization order needs to be increased, which will increase the transmission volume and system complexity. Compared with traditional receiving devices, this invention can achieve a good trade-off between the quantization order and the pulse compression effect. On the one hand, using single-bit quantization can minimize the complexity of quantization to the greatest extent. On the other hand, using a sidelobe suppression method based on a convolutional window function can effectively solve the problem of a significant decrease in the main lobe-to-side lobe ratio caused by the reduction of the quantization order.

[0053] (2) Compared with other sidelobe suppression techniques, the pulse compression sidelobe suppression method provided by this invention uses a convolution window function method with low computational complexity and high flexibility. It can flexibly adjust the type and order of the window function according to specific application requirements and the spectral characteristics of different window functions, thereby achieving different main lobe widths and main-side lobe ratios.

[0054] (3) Compared with traditional broadband digital receivers, the pulse compression sidelobe suppression method provided by this invention uses single-bit quantization technology, which improves the operation speed of broadband digital receivers, reduces system resource consumption and complexity, and reduces the overall size of the receiver by reducing the computational amount of discrete Fourier transform.

[0055] (4) For radar signals with large time-bandwidth products, such as linear frequency modulated signals, the pulse compression sidelobe suppression method provided by this invention can improve the pulse compression effect, increase the radar's effective range, and at the same time ensure good range resolution. It has certain reference value for improving the design in radar engineering technology.

[0056] (5) This invention can be applied to the sidelobe suppression of high-resolution synthetic aperture lidar images. If there is a high sidelobe level in the image, the image will become blurry and the quality will drop significantly. This method can suppress the sidelobe to a certain extent and avoid the main lobe from widening, thereby setting the excess data to zero, effectively reducing the amount of data, and thus improving the image contrast.

[0057] (6) This invention can be applied to synthetic aperture radar imaging methods based on single-bit quantization, and can further improve the problem of image quality degradation caused by single-bit quantization. Numerous analytical studies have shown that although single-bit quantization changes the signal waveform and broadens the signal spectrum, under conditions of low signal-to-noise ratio and oversampling, it is still possible to obtain imaging performance comparable to that of high-bit quantization using traditional matched filtering methods with single-bit data. This invention can improve the deterioration of traditional matched filtering effects.

[0058] (7) This invention can be applied to the detection and tracking of small targets under strong clutter, improve the contradiction between the radar's effective range and range resolution, improve the radar's range resolution, and is an effective way to achieve high radar resolution. It is also a powerful means for radar to counter stealth, resist electronic interference, and counter anti-radiation missiles.

[0059] (8) The present invention obtains a pulse-compressed received signal by multiplying the new matched filter signal with the quantized Chirp echo signal, which can improve the range resolution of radar target detection and significantly reduce false alarms and missed alarms.

[0060] (9) This invention can be extended to the processing of radar signals in other scenarios, and has certain reference value for subsequent engineering practice. Attached Figure Description

[0061] Figure 1 This is a block diagram illustrating the single-bit quantization principle of the present invention;

[0062] Figure 2 This is a flowchart of the method of the present invention;

[0063] Figure 3 This invention is the result of pulse compression of Chirp echo signals before and after single-bit quantization.

[0064] Figure 4 The second-, third-, and fourth-order self-convolution spectral characteristics of the rectangular window of this invention.

[0065] Figure 5 The second-, third-, and fourth-order self-convolution spectral characteristics of the Hamming window of this invention.

[0066] Figure 6 These are the spectral characteristics of the rectangular window and Hamming window of this invention, as well as the spectral characteristics of the second-order hybrid convolution.

[0067] Figure 7 This is a comparison of pulse compression results before and after the processing of this invention. Detailed Implementation

[0068] This invention provides a single-bit quantization chirp signal pulse compression sidelobe suppression method based on a convolutional window function. On the one hand, single-bit quantization can alleviate the burden on high-speed ADCs and reduce the total number of data bits, thereby reducing the burden of data storage and transmission. On the other hand, the convolutional window function can solve the problem of decreased main lobe-to-sidelobe ratio caused by single-bit quantization, avoiding false alarms or missed detections of small targets near the sidelobes, which has certain reference value for subsequent engineering practice.

[0069] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0070] like Figure 2 As shown in the figure, the pulse compression sidelobe suppression method for a single-bit quantized Chirp signal provided in this embodiment includes the following steps:

[0071] Step S1: Generate Chirp echo signal;

[0072] Step S2: Perform single-bit quantization on the Chirp echo signal to obtain the quantized Chirp echo signal;

[0073] Step S3: Perform single-bit quantization on the conjugate matched filter signal of the Chirp echo signal and add a convolution window to obtain a new matched filter signal;

[0074] Step S4: Use a new matched filter signal to perform pulse compression on the quantized Chirp echo signal to obtain the received signal.

[0075] The Chirp echo signal is quantized by a single bit to obtain the quantized signal. Specifically, the I and Q channels of the Chirp echo signal are encoded by a comparator. If the original value of the I or Q channel signal is positive, it is quantized to the value 1. If the original value is negative, it is quantized to the value -1.

[0076] The conjugate matched-filtered signal of the Chirp echo is quantized by a single bit and then a convolution window is applied to obtain a new matched-filtered signal, specifically as follows:

[0077] Step S21: Perform single-bit quantization on the conjugate matched filter signal of the original echo signal to obtain the quantized conjugate matched filter signal;

[0078] Step S22: Based on the frequency domain characteristics of various window functions and the requirements of actual application scenarios, the simulation determines the type and number of window functions participating in convolution, and obtains the selected convolution window functions.

[0079] Step S23: Apply the obtained convolution window function to the quantized conjugate matched filter signal to truncate it, and obtain a new matched filter signal.

[0080] Step S24 uses a new matched filter signal to perform pulse compression on the quantized Chirp echo signal to obtain the received signal. Specifically, the new matched filter signal is multiplied by the quantized Chirp echo signal to obtain the pulse-compressed received signal.

[0081] In step S22, based on the frequency domain characteristics of various window functions and the requirements of actual application scenarios, the simulation determines the type and number of window functions participating in convolution, thus obtaining the selected convolution window functions, specifically:

[0082] Step S221: Calculate the spectrum function of the cosine window;

[0083] Step S222: Calculate the frequency domain characteristics of the first-order single window function based on the spectrum function, and determine whether the frequency domain characteristics meet the requirements. If not, proceed to step S3; if yes, obtain the selected convolution window function.

[0084] Step S223: Calculate the frequency domain characteristics of the Nth-order single window function and determine whether the frequency domain characteristics meet the requirements. If not, proceed to step S4; if yes, obtain the selected convolution window function.

[0085] Step S224: Calculate the frequency domain characteristics of the mixed convolution of rectangular window with triangular window, Hanning window, Hamming window or Blackman window; determine whether the frequency domain characteristics meet the requirements. If not, proceed to step S5; if yes, obtain the selected convolution window function.

[0086] Step S225: Calculate the frequency domain characteristics of the mixed convolution of a first-order rectangular window with an Nth-order triangular window, Hanning window, Hamming window, or Blackman window; select the spectral characteristics that best meet the requirements, and the corresponding convolution window is the selected convolution window function; where N is an integer less than or equal to 4.

[0087] The specific method for calculating the spectral function of a cosine window is as follows:

[0088]

[0089] In the formula, L is the number of cosine terms, al is the coefficient of the cosine combination term, M is the length of the cosine window, and W R The expression for calculating (w) is:

[0090] The frequency domain characteristics of a first-order single window function are as follows:

[0091] Rectangular window: Window length is N, main lobe width is 4π / N, and side lobe peak value is less than -13.31dB;

[0092] Triangular window: Window length is N, main lobe width is 8π / N, and side lobe peak value is less than -26.63dB;

[0093] Hanning window: Window length is N, main lobe width is 8π / N, and side lobe peak value is less than -31.5dB;

[0094] Hamming window: Window length is N, main lobe width is 8π / N, and side lobe peak value is less than -44.7dB;

[0095] Blackman window: window length is N, main lobe width is 12π / N, and side lobe peak value is less than -58.22dB.

[0096] The frequency domain characteristics of an Nth-order single window function are as follows:

[0097] Rectangular window: When the order is 2 to 4, the window length is N, the main lobe width is 4π / N, and the side lobe peak value is -53.25 to -26.62 dB;

[0098] For triangular windows of order 2 to 4, the window length is N, the main lobe width is 8π / N, and the side lobe peak value is -106.6 to -53.25 dB.

[0099] Hanning window: When the order is 2 to 4, the window length is N, the main lobe width is 8π / N, and the side lobe peak value is -126 to -63 dB;

[0100] Hamming window: When the order is 2 to 4, the window length is N, the main lobe width is 8π / N, and the side lobe peak value is -178.8 to -89.41 dB;

[0101] Blackman window: When the order is 2 to 4, the window length is N, the main lobe width is 12π / N, and the side lobe peak value is -232.9 to -174.7 dB.

[0102] The frequency domain characteristics of mixed convolutions of rectangular windows with triangular windows, Hanning windows, Hamming windows, or Blackman windows are as follows:

[0103] Using a hybrid convolution with rectangular and triangular windows, the sidelobe peak value is less than -28.34dB, and the main lobe width is 4π / N;

[0104] Using a hybrid convolution of rectangular and Hanning windows, the sidelobe peak value is less than -24.96dB, and the main lobe width is 4π / N;

[0105] Using a hybrid convolution of rectangular and Hamming windows, the sidelobe peak value is less than -27.54dB, and the main lobe width is 4π / N;

[0106] Using a hybrid convolution of rectangular and Blackman windows, the sidelobe peak value is less than -21.83dB, and the main lobe width is 4π / N.

[0107] The frequency domain characteristics of the mixed convolution of a first-order rectangular window with an Nth-order triangular window, Hanning window, Hamming window, or Blackman window are as follows: the pulse compression result of the mixed convolution of a first-order rectangular window and an Nth-order triangular window is to suppress the sidelobes to -28.34dB to -62.06dB; the pulse compression result of the mixed convolution of a first-order rectangular window and an Nth-order Hanning window is to suppress the sidelobes to -24.96dB to -51.01dB; the pulse compression result of the mixed convolution of a first-order rectangular window and an Nth-order Hamming window is to suppress the sidelobes to -27.54dB to -58.92dB; and the pulse compression result of the mixed convolution of a first-order rectangular window and an Nth-order Blackman window is to suppress the sidelobes to -21.83dB to -42dB; where N is an integer less than or equal to 4.

[0108] Example

[0109] The single-bit quantization method designed in this example is as follows: Figure 1 As shown, after obtaining the chirp echo signal, the I and Q channels of the echo signal are encoded using single-bit encoding. This process is implemented using a comparator. If the original value of the I channel signal is positive, it is quantized to the value 1; if the value is negative, it is quantized to the value -1. The quantization method for the Q channel signal is similar. In this way, by mapping the original real-valued I and Q channels to two discrete values, the original echo signal is quantized into {±1,±j}, thus obtaining the single-bit quantized signal of the original echo signal. This quantization method reduces the burden of data storage and transmission by reducing the quantization bit precision of the quantizer, thereby reducing the complexity and cost of hardware design.

[0110] A convolutional window refers to the temporal convolution of multiple identical or different cosine windows. When multiple identical cosine windows are convolved, it is called a cosine self-convolutional window. When multiple different cosine windows are convolved, it is called a hybrid convolutional window or a combined convolutional window. The frequency domain performance of a convolutional window is closely related to the number and coefficients of the cosine combination terms. The main purpose of this embodiment is to construct a combined window function with low sidelobe peak values ​​and fast sidelobe decay.

[0111] The typical time-domain expression for a cosine window is:

[0112]

[0113] In the above formula, L is the number of cosine terms, m = 0, 1, 2, ..., M-1, M is the length of the window, and a l Here is the coefficient of the cosine combination term, a. l The following constraints must be met:

[0114]

[0115] The spectrum function of the cosine window is:

[0116]

[0117] in

[0118] A cosine self-convolution window can be obtained by performing temporal self-convolution on a cosine window:

[0119]

[0120] In the formula, p represents the number of cosine windows in the self-convolution operation, i.e., the order of the cosine self-convolution window. Let the length of x(m) be M, and from the properties of convolution, we get x c A sequence of length pM-p+1 (n) is given by x. c (n) Padding with p-1 zeros yields a sequence x of length pM. c (n). The process of generating a hybrid convolutional window is similar.

[0121] The selection strategy for single window function types is as follows: Rectangular windows have the narrowest main lobe width but higher side lobe peaks; compared to rectangular windows, Hamming and Hanning windows have wider and lower main lobes, while their side lobe amplitudes are significantly reduced, resulting in better performance than rectangular windows, but the increased main lobe width leads to a decrease in resolution; compared to rectangular, triangular, Hamming, and Hanning windows, Blackman windows have the widest main lobe width but also lower side lobe peaks. The table below summarizes the frequency domain characteristics of commonly used window functions.

[0122] Table 1: Frequency Domain Characteristics of Several Classical Window Functions

[0123] Window function Window length Main lobe width Side lobe peak value / dB rectangular window N 4π / N -13.31 triangular window N 8π / N -26.63 Hanning Window N 8π / N -31.5 Haiming Window N 8π / N -44.7 Blackman window N 12π / N -58.22

[0124] The strategy for selecting the order of a single window function is as follows: the main lobe width of a p-order cosine self-convolution window serves as an indicator of its frequency resolution capability. Its value is determined by the length of the original window function participating in the self-convolution operation and is inversely proportional to the length of the original window function. The p-value of the p-order cosine self-convolution window is positively correlated with its ability to suppress spectral leakage; that is, the larger the p value, the lower the sidelobes and the stronger the ability to suppress spectral leakage. The self-convolution function optimizes the sidelobe performance of the cosine self-convolution window, thereby improving its ability to suppress spectral leakage. The table below summarizes the frequency domain characteristics of commonly used window functions at different orders.

[0125] Table 2: Frequency domain characteristics of several classic window functions at different orders

[0126]

[0127]

[0128] The strategy for selecting hybrid convolutional window functions is as follows: Generally, a single window function cannot meet the needs of practical applications; multiple windows need to be combined in convolution to achieve better windowing results. The main lobe width of a hybrid convolutional window is the smallest among the participating window functions, and the side lobe peak value lies between the largest and smallest side lobe peak values. This approach can take into account the advantages of all participating window functions. Therefore, a compromise can be achieved by selecting a window function with a narrow main lobe width and a window function with a low side lobe peak value for combined convolution. Rectangular window functions have the narrowest main lobe width. To simultaneously obtain a narrow main lobe width and a high main-to-side lobe ratio, rectangular windows are often combined with other types of window functions in convolution. The table below summarizes the frequency domain characteristics of rectangular windows in convolution with other types of window functions.

[0129] Table 3: Frequency domain characteristics of convolution with rectangular window and other types of window functions

[0130] Window function Window length Main lobe width Side lobe peak value / dB Rectangular windows and triangular windows N 4π / N -28.34 Rectangular windows and Hanning windows N 4π / N -24.96 Rectangular windows and Hamming windows N 4π / N -27.54 Rectangular windows and Blackman windows N 4π / N -21.83

[0131] The strategy for selecting the order of the hybrid convolutional window function is as follows: the sidelobe performance of the hybrid convolutional window is continuously optimized as the convolution order increases, while ensuring that the main lobe width is the smallest among the window functions participating in the convolution. Taking the hybrid convolution of a rectangular window with a triangular window, a Hanning window, a Hamming window, and a Blackman window as examples, the spectral performance of their second to fifth-order hybrid convolutional windows is analyzed, as shown in the table below.

[0132] Table 4: Frequency Domain Characteristics of Hybrid Convolution of Rectangular Windows and Other Window Types

[0133] Window function Window length Main lobe width Side lobe peak value / dB Rectangular windows (level 1) and triangular windows (levels 1-4) N 4π / N -28.34～-62.06 Rectangular window (level 1) and Hanning window (levels 1-4) N 4π / N -24.96～-51.01 Rectangular windows (level 1) and Hamming windows (levels 1-4) N 4π / N -27.54～-58.92 Rectangular window (level 1) and Blackman window (levels 1-4) N 4π / N -21.83～-42

[0134] When multiple targets are detected, the main lobe of a nearby small target signal can be overwhelmed by the side lobes of a larger target signal, causing the small target to be undetectable and resulting in target loss. This false alarm situation is mainly improved by suppressing pulse compression sidelobes, thereby increasing the main-to-side lobe ratio. Therefore, the sidelobe peak value is often used to evaluate the suppression effect. A hybrid convolution of rectangular and triangular windows can suppress the sidelobes of the pulse compression result to -28.34 dB to -62.06 dB; a hybrid convolution of rectangular and Hanning windows can suppress them to -24.96 dB to -51.01 dB; a hybrid convolution of rectangular and Hamming windows can suppress them to -27.54 dB to -58.92 dB; and a hybrid convolution of rectangular and Blackman windows can suppress them to -21.83 dB to -42 dB. This can significantly reduce false alarms and missed alarms.

[0135] In pulse compression systems, the transmitted signal bandwidth is often very large. The receiving system compresses the echo signal, and the bandwidth after compression is B. Then, the range resolution after pulse compression can be expressed as:

[0136]

[0137] Therefore, the narrower the main lobe width after pulse compression, the higher the range resolution of the system. By adding a hybrid convolution window, the main lobe width can be compressed to 4π / N, and the range resolution can be improved to cN / 8π, where N is the length of the window function; this can significantly improve the range resolution of radar target detection.

[0138] The results of pulse compression of the Chirp echo signals before and after single-bit quantization are as follows: Figure 3 As shown, single-bit quantization of the echo signal significantly affects the waveform of the received pulse compression signal, and the spectrum of the received pulse compression signal is correspondingly broadened. Single-bit quantization significantly degrades the performance of matched reception, increases the sidelobes of the pulse compression signal, and decreases the main-to-side-lobe ratio. This will seriously affect the detection of small targets near the sidelobes or cause false targets, which is not conducive to practical engineering applications. Therefore, the sidelobe suppression problem of the pulse compression signal after single-bit quantization must be considered.

[0139] This example compares the spectral characteristics of second-, third-, and fourth-order self-convolution functions of rectangular and Hamming windows when selecting the type and order of window functions, as shown below. Figure 4 and Figure 5 As shown in the figure, it can be seen that the sidelobe performance of the cosine self-convolution window can be continuously optimized as the convolution order increases. The main lobe width after convolution is equal to the main lobe width of the original window function. Therefore, the self-convolution window can reduce the sidelobe peak value while keeping the main lobe width unchanged. However, the performance of a single window function cannot meet the actual needs at this time. Therefore, this embodiment selects a hybrid convolution window function. The designed hybrid convolution window is a second-order hybrid convolution of a rectangular window function and a Hamming window function. The spectra of the rectangular window and the Hamming window and the spectral characteristics of the second-order hybrid convolution are shown in the figure. Figure 6 As shown, the main lobe width of the second-order hybrid convolutional window is the smallest among the participating window functions (i.e., rectangular and Hamming windows), and the side lobe peaks are located between the largest and smallest side lobe peaks among the participating window functions (i.e., rectangular and Hamming windows). Hybrid convolutional windows can combine the advantages of the participating window functions. For example, when a rectangular window is convolved with a Hamming window, the main lobe width after convolution is equal to the main lobe width of the narrowest participating window, and the side lobe peaks are also lower than those of the narrowest window. The difficulty of this step lies in selecting the appropriate window function type and order for the convolution based on the specific application requirements and the frequency domain performance of various window functions.

[0140] The method of this invention performs matched filtering on the single-bit quantized chirp echo signal. By performing single-bit quantization on the conjugate matched filter signal of the echo signal and adding a second-order hybrid convolution window for truncation, the pulse compression sidelobes of the single-bit quantized chirp signal can be effectively suppressed. A comparison of the effects before and after pulse compression is provided. Figure 7 As shown. Figure 7 The red curve represents the unprocessed pulse compression output, with a main-to-side lobe ratio of approximately 5.9 dB. In general, modern radar pulse compression performance requires a main-to-side lobe ratio of over 30 dB, which is completely insufficient for the actual operation of radar. Figure 7 The blue curve represents the pulse compression output after processing using this method. It can be seen that the main-to-side-lobe ratio of the pulse compression output is significantly reduced to approximately -30.47 dB after processing, with little change in the main lobe width, which meets the operational requirements of practical radar. Verification by examples shows that the pulse compression sidelobe suppression method for single-bit quantized chirp signals based on hybrid convolutional window functions provided in this invention can effectively solve the problem of pulse compression main-to-side-lobe ratio degradation caused by single-bit quantization technology. The sidelobe suppression effect is significant, improving the radar's range resolution.

[0141] Although the present invention has been described in detail through the preferred embodiments above, it should be understood that the above description should not be considered as a limitation of the present invention. Various modifications and substitutions to the present invention will be apparent to those skilled in the art after reading the above description. Therefore, the scope of protection of the present invention should be defined by the appended claims.

[0142] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A method for suppressing pulse compression sidelobes in a single-bit quantized Chirp signal, characterized in that, include: Generate Chirp echo signal; The Chirp echo signal is quantized using a single bit to obtain the quantized Chirp echo signal; The conjugate matched filter signal of the Chirp echo signal is quantized by a single bit and a convolution window is added to obtain a new matched filter signal; The quantized Chirp echo signal is pulse-compressed using a new matched filter signal to obtain the received signal. The process of performing single-bit quantization on the conjugate matched filter signal of the Chirp echo signal and adding a convolution window to obtain a new matched filter signal is as follows: The conjugate matched filter signal of the original echo signal is quantized by a single bit to obtain the quantized conjugate matched filter signal; Based on the frequency domain characteristics of various window functions and the requirements of actual application scenarios, simulation is used to determine the type and number of window functions participating in convolution, thus obtaining the selected convolution window functions. The obtained convolutional window function is used to window and truncate the quantized conjugate matched filter signal to obtain a new matched filter signal. Based on the frequency domain characteristics of various window functions and the requirements of actual application scenarios, the simulation determines the type and number of window functions participating in convolution, thus obtaining the selected convolution window functions, specifically: S1. Calculate the spectral function of the cosine window; S2. Calculate the frequency domain characteristics of the first-order single window function based on the spectrum function, and determine whether the frequency domain characteristics meet the requirements. If not, proceed to step S3; if yes, obtain the selected convolution window function. S3. Calculate the frequency domain characteristics of the Nth-order single window function and determine whether the frequency domain characteristics meet the requirements. If not, proceed to step S4; if yes, obtain the selected convolution window function. S4. Calculate the frequency domain characteristics of the mixed convolution of rectangular window with triangular window, Hanning window, Hamming window or Blackman window; determine whether the frequency domain characteristics meet the requirements. If not, proceed to step S5; if yes, obtain the selected convolution window function. S5. Calculate the frequency domain characteristics of the mixed convolution of a first-order rectangular window with an Nth-order triangular window, Hanning window, Hamming window, or Blackman window; select the spectral characteristics that best meet the requirements, and the corresponding convolution window is the selected convolution window function; where N is an integer less than or equal to 4.

2. The method for suppressing pulse compression sidelobes of a single-bit quantized Chirp signal according to claim 1, characterized in that, The process of performing single-bit quantization on the Chirp echo signal to obtain the quantized signal is as follows: the I and Q channels of the Chirp echo signal are encoded using a comparator. If the original value of the I or Q channel signal is positive, it is quantized to the value 1; if the original value is negative, it is quantized to the value -1.

3. The method for suppressing pulse compression sidelobes of a single-bit quantized Chirp signal according to claim 1, characterized in that: The method of using a new matched filter signal to pulse compress the quantized Chirp echo signal to obtain the received signal is as follows: the new matched filter signal is multiplied by the quantized Chirp echo signal to obtain the pulse-compressed received signal.

4. The pulse compression sidelobe suppression method for a single-bit quantized Chirp signal according to claim 1, characterized in that: The specific method for calculating the spectral function of the cosine window is as follows: In the formula, Let cosine be the number of terms. The coefficients of the cosine combination term are... The length of the cosine window.

5. The method for suppressing pulse compression sidelobes of a single-bit quantized Chirp signal according to claim 1, characterized in that: The frequency domain characteristics of the first-order single window function are as follows: Rectangular window: Window length is N, main flap width is... The sidelobe peak value is less than -13.31 dB; triangular window: Window length is N, main flap width is... The sidelobe peak value is less than -26.63 dB; Hanning window: Window length is N, main petal width is The sidelobe peak value is less than -31.5 dB; Hamming window: Window length is N, main lobe width is... The sidelobe peak value is less than -44.7 dB; Blackman window: Window length is N, main lobe width is The sidelobe peak value is less than -58.22dB.

6. The method for suppressing pulse compression sidelobes of a single-bit quantized Chirp signal according to claim 1, characterized in that: The frequency domain characteristics of the Nth-order single window function are as follows: Rectangular window: When the order is 2~4, the window length is N, and the main lobe width is... The sidelobe peak value is -53.25 to -26.62 dB; Triangular window: When the order is 2~4, the window length is N, and the main lobe width is... The sidelobe peak value is -106.6 to -53.25 dB; Hanning window: When the order is 2-4, the window length is N, and the main lobe width is... The sidelobe peak value is -126 to -63 dB; Hamming window: When the order is 2-4, the window length is N, and the main lobe width is... The sidelobe peak value is -178.8 to -89.41 dB; Blackman window: When the order is 2-4, the window length is N, and the main lobe width is... The sidelobe peak value is -232.9 to -174.7 dB.

7. The method for suppressing pulse compression sidelobes of a single-bit quantized Chirp signal according to claim 1, characterized in that: The frequency domain characteristics of the mixed convolution of the rectangular window with the triangular window, Hanning window, Hamming window, or Blackman window are as follows: Using a hybrid convolution method employing rectangular and triangular windows, the sidelobe peak value is less than -28.34 dB, and the main lobe width is [missing value]. ; Using a hybrid convolution method combining rectangular and Hanning windows, the sidelobe peak value is less than -24.96 dB, and the main lobe width is [missing value]. ; Using a hybrid convolution method combining rectangular and Hamming windows, the sidelobe peak value is less than -27.54 dB, and the main lobe width is [missing value]. ; Using a hybrid convolution method combining rectangular and Blackman windows, the sidelobe peak value is less than -21.83 dB, and the main lobe width is [missing value]. .

8. The method for suppressing pulse compression sidelobes of a single-bit quantized Chirp signal according to claim 1, characterized in that: The frequency domain characteristics of the mixed convolution of the first-order rectangular window with the Nth-order triangular window, Hanning window, Hamming window, or Blackman window are as follows: The hybrid convolution pulse compression results of first-order rectangular window and Nth-order triangular window suppress sidelobes to -28.34 dB ~ -62.06 dB; The hybrid convolution pulse compression results of first-order rectangular window and N-order Hanning window suppress sidelobes to -24.96 dB ~ -51.01 dB; The hybrid convolution pulse compression results of first-order rectangular window and N-order Hamming window suppress sidelobes to -27.54 dB ~ -58.92 dB; The hybrid convolution pulse compression results of first-order rectangular window and N-order Blackman window suppress sidelobes to -21.83 dB ~ -42 dB; Where N is an integer less than or equal to 4.

Images