A method for generating a nonlinear frequency modulation signal in real time based on ZYNQ
By calculating and transmitting frequency and phase control words in real time to the DDS module at the PL end through the PS end of the ZYNQ chip, the problem of wasted NLFM signal resources is solved, and flexible and efficient NLFM signal generation is realized.
Patent Information
- Application Number
- CN202410682799.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-29
- Publication Date
- 2026-01-13
- Estimated Expiration
- 2044-05-29
AI Technical Summary
In existing technologies, generating NLFM signals requires pre-generated data to be stored in the FPGA's ROM, resulting in wasted resources and insufficient flexibility.
The frequency and phase control words are calculated at the PS end of the ZYNQ chip and transmitted in real time to the DDS module at the PL end via the AXI4-Lite protocol to directly generate NLFM signals, eliminating the need for ROM to store the frequency and phase control words.
It saves FPGA resources, improves the flexibility and adjustability of generating NLFM signals, and is simple to operate without requiring pre-calculation by a host computer.
Smart Images

Figure CN118655528B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of radar signal generation technology, specifically relating to a method for generating nonlinear frequency modulated signals in real time based on ZYNQ. Background Technology
[0002] Synthetic Aperture Radar (SAR) is an advanced radar technology that acquires radar images of surface targets by transmitting continuous radio pulses and receiving and processing the returned signals. Linear Frequency Modulation (LFM) is a common radar transmitted waveform. Matching and filtering this signal at the receiver results in a low main-to-side lobe ratio and a wide main lobe width, leading to azimuth and range ambiguity. Using Nonlinear Frequency Modulation (NLFM) can improve the clarity and accuracy of radar images.
[0003] Compared to LFM signals, NLFM signals can achieve a high main-to-side lobe ratio and a narrow main lobe width by directly performing matched filtering at the receiver. NLFM can effectively reduce azimuth ambiguity and improve image azimuth resolution through nonlinear frequency modulation. NLFM can also reduce range ambiguity, making the target's range and position in the radar image clearer and more accurate. NLFM signals are also more resistant to interference during transmission and reception, making the performance of NLFM radar systems more stable and reliable in complex electromagnetic environments.
[0004] Currently, methods for generating NLFM signals using FPGAs all employ the DDS+ROM approach. The host computer software calculates relevant integrals and inverse functions to determine the frequency and phase of the NLFM signal, which are then stored in the FPGA's ROM. The DDS then reads the frequency and phase data from the ROM and transmits the determined signal. This method has a significant drawback: when different signals need to be transmitted, data needs to be pre-generated and stored in the ROM. Due to the limited resources of the FPGA, the frequency and phase data consume a large amount of FPGA resources, leading to resource waste.
[0005] Therefore, how to conserve resources for generating NLFM signals and improve the flexibility and adjustability of generating new NLFM signals are pressing problems that need to be solved. Summary of the Invention
[0006] The purpose of this invention is to provide a method for generating nonlinear frequency modulated signals in real time based on ZYNQ, so as to solve the problem that generating NLFM signals requires rewriting data into ROM, which results in high resource consumption.
[0007] To achieve the above objectives, the present invention adopts the following technical solution:
[0008] A method for generating nonlinear frequency modulated signals in real time based on ZYNQ includes the steps of calculating the frequency control word and phase control word at the PS end, and completing the phase conversion at the PL end to obtain the output signal;
[0009] The calculation method for the PS terminal frequency control word and phase control word includes the following steps:
[0010] Step S1: Obtain the bandwidth, duration, window function type, and number of frequency points of the NLFM signal;
[0011] Step S2: Construct the group delay function T(f) based on the bandwidth, duration, window function type, and number of frequency points obtained in S1; calculate the time difference between adjacent frequency points based on the duration and number of frequency points obtained in S1 to determine the timer loading value;
[0012] Step S3: Determine the frequency and phase data of the NLFM signal based on the group delay function T(f);
[0013] Step S4: Convert the frequency data and phase data obtained in step S3 into frequency control words and phase control words;
[0014] Step S5: Determine whether the timer interrupt has reached the timer load value determined in step S2. If yes, transmit the frequency control word and phase control word to PL via the AXI4-Lite protocol; otherwise, return to step S3 to redetermine the frequency and phase of the NLFM signal.
[0015] The method for obtaining an output signal by completing phase conversion at the PL terminal includes the following steps:
[0016] The received frequency control word and phase control word are used to obtain the address of the ROM table through the DDS module. The corresponding amplitude is found by looking up the ROM table, and the phase to amplitude conversion is completed to obtain the output signal.
[0017] Furthermore, the group delay function T(f) obtained in step S2 is:
[0018]
[0019]
[0020] Where K represents a constant coefficient, B is the bandwidth of the NLFM signal, W(x) is the window function, f represents the frequency, and T represents the duration.
[0021] The formula for calculating the time difference between adjacent frequency points is:
[0022]
[0023] Among them, f psis the system clock value at the PS terminal, and x is the timer load value at the output.
[0024] Furthermore, step S3, which determines the frequency and phase of the NLFM signal based on the group delay function T(f), includes the following steps:
[0025] Taking the inverse of the group delay function T(f), we obtain the frequency modulation function f(t) of the NLFM signal: f(t) = T -1 (f), 0 ≤ t ≤ T; where T -1 (f) represents finding the inverse function of T(f);
[0026] The frequency data of the NLFM signal can be determined based on the frequency modulation function f(t);
[0027] The frequency modulation function f(t) of the NLFM signal is integrated using the integral approximation rectangular method to obtain the phase θ of the NLFM signal. i : Where, θ i f represents the phase data output at point i. out_i This represents the frequency data output at point i, which is the value of the inverse function. This represents the step size of the rectangle.
[0028] Furthermore, step S3 employs numerical analysis to find the inverse function of the group delay function T(f); and uses Newton's iteration method to determine the frequency data of the NLFM signal based on the frequency modulation function f(t).
[0029] Furthermore, the formula for converting the frequency data and phase data into frequency control words and phase control words in step S4 is as follows:
[0030]
[0031] Where K1 represents the frequency control word, K2 represents the phase control word, fout and θout represent the frequency and phase data to be output, respectively, N1 represents the number of bits of the phase accumulator, and N2 represents the number of bits of the waveform data table address.
[0032] This invention provides a method for real-time generation of nonlinear frequency modulated (NLFM) signals based on ZYNQ. This method achieves accurate and rapid calculation of each frequency and phase by inverting the group delay function at the PS end and performing subsequent integration operations, thus ensuring that different input parameters yield different NLFM signals without requiring pre-calculation by a host computer. Since the AXI4-Lite protocol between the PS and PL ends of the ZYNQ chip ensures high-speed data transmission, transmitting the frequency and phase control words calculated at the PS end to the DDS module at the PL end via the AXI4-Lite protocol eliminates the need for ROM storage of these control words, significantly saving resources. Compared to traditional methods for generating NLFM signals, this invention's NLFM signal generation method directly obtains the NLFM signal parameters used for frequency and phase calculations in the ZYNQ via a serial port, without requiring recompilation. This provides greater flexibility and adjustability, and the operation steps are simple. Attached Figure Description
[0033] Figure 1 This is a flowchart of a method for generating a nonlinear frequency modulated signal based on ZYNQ according to the present invention;
[0034] Figure 2 This is a flowchart illustrating how the inverse function of the group delay function is obtained using Newton's iteration method to determine the next frequency data in an embodiment.
[0035] Figure 3 This is a block diagram of the DDS at the PL end in the embodiment;
[0036] Figure 4 This is a schematic diagram of the AXI4-Lite protocol waveform in the embodiment where the PS transmits the frequency control word and phase control word to the PL;
[0037] Figure 5 These are the simulation time-domain curves of NLFM in the embodiments;
[0038] Figure 6 This is a comparison chart of the pulse compression outputs of the ideal NLFM signal and the simulated NLFM signal in the embodiment;
[0039] Figure 7 This is a comparison chart of the spectrum of the ideal NLFM signal and the simulated NLFM signal in the embodiment. Detailed Implementation
[0040] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.
[0041] like Figure 1 The present embodiment provides a method for generating nonlinear frequency modulated signals in real time based on ZYNQ, which includes the steps of calculating the frequency control word and the phase control word at the PS end, and completing the phase conversion at the PL end to obtain the output signal.
[0042] The calculation method for the PS terminal frequency control word and phase control word includes the following steps:
[0043] Step S1: Obtain the bandwidth, duration, window function type, and number of frequency points of the NLFM signal. In this embodiment, the bandwidth, duration, window function type, and number of frequency points of the NLFM signal are obtained through the serial port on the PS end. Specifically:
[0044] The serial port sends 12 hexadecimal data entries at a time. The first four hexadecimal data entries represent the bandwidth of the NLFM signal, the fifth to eighth represent the duration of the NLFM signal, the ninth to eleventh represent the number of frequency points in the NLFM signal, and the twelfth hexadecimal data entry represents the window function type. Window function types include Hanning window, Hamming window, and Blackman window. The acknowledgment type of the acquired hexadecimal data is: 0 for Hanning window, 1 for Hamming window, 2 for Blackman window, and 3 to 15 for invalid data.
[0045] Assuming the window function type obtained in this embodiment is a Blackman window, the expression for the Blackman window is:
[0046]
[0047] Where B is the bandwidth of the NLFM signal, and f represents the frequency.
[0048] Step S2: Construct the group delay function T(f) based on the bandwidth, duration, window function type and number of frequency points obtained in S1; calculate the time difference between adjacent frequency points according to formula (6) based on the duration and number of frequency points obtained in S1, so as to determine the timer loading value.
[0049] In this embodiment, the group delay function T(f) is defined as follows:
[0050]
[0051]
[0052] Where K represents a constant coefficient, B is the bandwidth of the NLFM signal, W(x) is the expression for the window function, f represents the frequency, and T represents the duration.
[0053] Substituting the expression for the Blackman window into the group delay function T(f), the group delay function T(f) is rewritten to obtain the expression for the group delay function based on the Blackman window in this embodiment (4):
[0054]
[0055] Differentiating equation (4) yields expression (5):
[0056]
[0057] Based on formula (2), the expression for formula (6) can be obtained according to the system clock value at the PS end, the duration of the NLFM signal, and the number of frequency points:
[0058]
[0059] Among them, f ps is the system clock value at the PS terminal, and x is the timer load value at the output.
[0060] Step S3: Determine the frequency and phase data of the NLFM signal based on the group delay function T(f). The implementation method is as follows:
[0061] (1) By taking the inverse of the group delay function T(f), the frequency modulation function f(t) of the NLFM signal is obtained:
[0062] f(t) = T -1 (f), 0 ≤ t ≤ T; (7)
[0063] Among them, T -1 (f) represents finding the inverse function of T(f).
[0064] Since the expression of the inverse function cannot be directly obtained from the PS side, this embodiment uses numerical analysis to obtain the inverse function of the group delay function to address this issue.
[0065] (2) Based on the frequency modulation function f(t), the value of the inverse function at a specific point is obtained by Newton's iteration method, and its expression is (8).
[0066]
[0067] The calculation process of Newton's iteration method is as follows: Figure 2 As shown:
[0068] First, express the group delay function T(f) in the form y = T(f), then express T... -1 (f) is represented as T(f)-y=0, then T′(f) is calculated, an initial approximation value f0 is set, and (8) is used for iteration until the required accuracy is met or the iteration limit is reached.
[0069] (3) The phase of the NLFM signal is obtained by integrating the frequency modulation function f(t). In this embodiment, the integral approximation rectangle method is used to obtain the phase of the NLFM signal, resulting in equation (9). The phase θ of the NLFM signal can be determined according to equation (9). i :
[0070]
[0071] Where, θ i f represents the phase data output at point i. out_i This represents the frequency data output at point i, which is the value of the inverse function. This represents the step size of the rectangle.
[0072] Step S4: Based on the working principle of DDS, the value of frequency control word K1 controls the frequency of the output signal, and the value of phase control word K2 controls the phase shift of the output signal, which is used for phase signal modulation. The frequency data and phase data obtained in step S3 are converted into frequency control words and phase control words; the conversion formula used in this implementation is Equation (10):
[0073]
[0074] Where K1 represents the frequency control word, K2 represents the phase control word, fout and θout represent the frequency and phase data to be output, respectively, f0 represents the DDS reference clock, N1 represents the number of bits of the phase accumulator, and N2 represents the number of bits of the waveform data table address.
[0075] Step S5: Determine whether the timer interrupt has reached the timer load value determined in step S2. If yes, transmit the frequency control word and phase control word to PL via the AXI4-Lite protocol; otherwise, return to step S3 to redetermine the frequency and phase of the NLFM signal.
[0076] In this embodiment, the waveform diagram of the PS transmitting the frequency control word and phase control word to the PL via the AXI4-Lite protocol is shown below. Figure 4 As shown. Specifically:
[0077] On the rising edge of the clock, the correct address is assigned to M_AXI_AWADDR and the M_AXI_AWVALID signal is pulled high. The system waits for M_AXI_AWREADY to go high to complete the first handshake. After M_AXI_AWREADY goes high, M_AXI_AWVALID is immediately pulled low, and the M_AXI_WVALID signal is pulled high along with the M_AXI_WDATA value. M_AXI_WDATA is 32 bits of data, with the first 20 bits being the frequency control word and the last 12 bits being the phase control word. If M_AXI_WREADY is detected to be high while M_AXI_WVALID is high, the second handshake is complete, and data transmission is finished.
[0078] The method for obtaining an output signal by completing phase conversion at the PL terminal includes the following steps:
[0079] The received frequency control word and phase control word are used to obtain the address of the ROM table through the DDS module. The corresponding amplitude is found by looking up the ROM table, and the phase to amplitude conversion is completed to obtain the output signal.
[0080] The schematic diagram of the DDS module structure at the PL end in this embodiment is as follows: Figure 3 As shown, the phase accumulator consists of an N1-bit adder and an N1-bit register. At the rising edge of each clock cycle, the adder adds the frequency control word to the phase data output from the accumulator register and feeds the result back to the data input of the accumulator register, allowing the adder to continue adding to the frequency control word under the next clock pulse. In this way, the phase accumulator continuously performs linear phase accumulation on the frequency control word under the clock pulse. That is, the phase accumulator accumulates the frequency control word once with each clock pulse input. The data output by the phase accumulator is the phase of the synthesized signal. The overflow frequency of the phase accumulator is the signal frequency output by the DDS. The data output by the phase accumulator is used as the phase sampling address of the waveform memory. The waveform sample value stored in the waveform memory can then be retrieved by looking up a table, thus completing the phase-to-amplitude conversion and obtaining the output signal.
[0081] Figure 5 These are the simulation time-domain curves of the NLFM implementation example. Figure 6 This is a comparison chart of the pulse compression outputs of the ideal NLFM signal and the simulated NLFM signal from the embodiment. Figure 7 This is a comparison chart of the spectrum of the ideal NLFM signal and the simulated NLFM signal from the embodiment. Figures 5-7 It can be seen that the nonlinear frequency modulation signal generation method of this embodiment is feasible.
[0082] It is understood that the present invention has been described through some embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.
Claims
1. A method for generating nonlinear frequency-modulated signals in real time based on ZYNQ, comprising two parts: calculating a frequency control word and a phase control word at the PS terminal, and performing phase conversion at the PL terminal to obtain the output signal, characterized in that: The calculation method for the PS terminal frequency control word and phase control word includes the following steps: Step S1: Obtain the bandwidth, duration, window function type, and number of frequency points of the NLFM signal; Step S2: Construct the group delay function T(f) based on the bandwidth, duration, window function type, and number of frequency points obtained in S1; calculate the time difference between adjacent frequency points based on the duration and number of frequency points obtained in S1 to determine the timer loading value; Step S3: Determine the frequency and phase data of the NLFM signal based on the group delay function T(f); Step S4: Convert the frequency data and phase data obtained in step S3 into frequency control words and phase control words; Step S5: Determine whether the timer interrupt has reached the timer load value determined in step S2. If yes, transmit the frequency control word and phase control word to PL via the AXI4-Lite protocol; otherwise, return to step S3 to redetermine the frequency and phase of the NLFM signal. The method for obtaining an output signal by completing phase conversion at the PL terminal includes the following steps: The received frequency control word and phase control word are used to obtain the address of the ROM table through the DDS module. The corresponding amplitude is found by looking up the ROM table, and the phase to amplitude conversion is completed to obtain the output signal.
2. The method for generating nonlinear frequency modulated signals in real time based on ZYNQ according to claim 1, characterized in that, The group delay function T(f) obtained in step S2 is: Where K represents a constant coefficient, B is the bandwidth of the NLFM signal, W(x) is the window function, f represents the frequency, and T represents the duration; The formula for calculating the time difference between adjacent frequency points is: Among them, f ps is the system clock value at the PS terminal, and x is the timer load value at the output.
3. The method for generating nonlinear frequency modulated signals in real time based on ZYNQ according to claim 2, characterized in that, The method for determining the frequency and phase of the NLFM signal based on the group delay function T(f) in step S3 includes the following steps: Taking the inverse of the group delay function T(f), we obtain the frequency modulation function f(t) of the NLFM signal: f(t) = T -1 (f), 0 ≤ t ≤ T; where T -1 (f) represents finding the inverse function of T(f); The frequency data of the NLFM signal can be determined based on the frequency modulation function f(t); The frequency modulation function f(t) of the NLFM signal is integrated using the integral approximation rectangular method to obtain the phase θ of the NLFM signal. i : Where, θ i f represents the phase data output at point i. out_i This represents the frequency data output at point i, which is the value of the inverse function. This represents the step size of the rectangle.
4. The method for generating nonlinear frequency modulated signals in real time based on ZYNQ according to claim 3, characterized in that: Step S3 uses numerical analysis to find the inverse function of the group delay function T(f); and Newton's iteration method is used to determine the frequency data of the NLFM signal based on the frequency modulation function f(t).
5. The method for generating nonlinear frequency modulated signals in real time based on ZYNQ according to claim 4, characterized in that: The formula for converting the frequency data and phase data in step S4 into frequency control words and phase control words is as follows: Where K1 represents the frequency control word, K2 represents the phase control word, and f out and θ out These represent the frequency and phase data to be output, respectively. N1 represents the number of bits in the phase accumulator, N2 represents the number of bits in the waveform data table address, and f0 represents the reference clock of the DDS.