A parameter adaptive real-time spectrum analysis method and system

Abstract

The application discloses a parameter self-adaptive real-time spectrum analysis method and system, and belongs to the technical field of digital signal processing.The method comprises the following steps: performing down-sampling filtering on an IQ signal by using a cascaded FIR filter based on a down-sampling coefficient to obtain a down-sampled IQ signal; dividing the down-sampled IQ signal into a non-over-sampling rate signal or an over-sampling rate signal according to an actual sampling rate, performing overlapping segmentation processing to obtain a serial or parallel segmented time domain signal, applying a window function with an actual FFT length and performing FFT conversion to obtain a plurality of groups of frequency domain signals, calculating the amplitudes of the plurality of groups of frequency domain signals and performing average processing to obtain a frequency-synthesized signal; performing logarithmic conversion on the frequency-synthesized signal by using a segmented look-up table method to obtain a logarithm-processed frequency domain signal, and performing video filtering on the logarithm-processed frequency domain signal to obtain a real-time spectrum signal.The application solves the problems of small bandwidth, long data processing time and complex structure existing in the prior art real-time spectrum analysis technology.

Description

Technical Field

[0001] This invention relates to a parameter-adaptive real-time spectrum analysis method and system, belonging to the field of digital signal processing technology. Background Technology

[0002] In modern wireless communication, radar detection, and electronic countermeasures, rapid and accurate measurement and monitoring of signal spectral characteristics are of paramount importance. Traditional spectrum analysis typically employs a superheterodyne spectrum analyzer, which can only calculate the amplitude data at one frequency point at a time. This necessitates lengthy scans when the sweep bandwidth is wide, and traditional spectrum analyzers may miss rapidly changing portions of the signal due to excessive sweep time. Real-time spectrum analyzers, however, combine the working principles of a swept superheterodyne spectrum analyzer and an FFT analyzer. Through a high-speed digital signal processing module, they acquire and analyze signals, storing discrete data and then using a Fast Fourier Transform (FFT) to derive the complete spectral data of the signal within the analysis bandwidth, thus ensuring no signal loss.

[0003] However, when signal bandwidth reaches the hundreds of megahertz to gigahertz levels, the conventional single-channel analog front-end plus digital down-conversion and single-channel Fast Fourier Transform (FFT) processing architecture has the following problems: The sampling rate supported by this architecture is limited by the clock frequency, and the sampling rate limits the bandwidth. It is difficult for a single-channel FFT to reach the gigahertz level sampling rate, thus limiting the bandwidth of spectrum analysis. Even if the architecture meets the high-speed sampling rate requirement, a single-channel FFT struggles to handle large amounts of data. When the processing time for one frame of data exceeds the data acquisition time, real-time signal processing becomes impossible. Therefore, conventional real-time spectrum analysis supports limited bandwidth and excessively long data processing times. While two-dimensional FFT algorithms can support large-bandwidth spectrum analysis and accelerate the processing of large amounts of data, the required resource overhead is too high and the structure is complex. In summary, existing real-time spectrum analysis technologies suffer from limited bandwidth, long data processing times, and complex structures. Summary of the Invention

[0004] The purpose of this invention is to provide a parameter-adaptive real-time spectrum analysis method and system, which solves the problems of small bandwidth, long data processing time and complex structure in existing real-time spectrum analysis technologies by adaptive parameter calculation and FPGA hardware optimization.

[0005] To solve the above-mentioned technical problems, the present invention is implemented using the following technical solution.

[0006] In a first aspect, the present invention provides a parameter-adaptive real-time spectrum analysis method, comprising:

[0007] Acquire the radio frequency signal to be analyzed;

[0008] The actual sampling rate, actual FFT length, and downsampling coefficient are calculated based on the start frequency, end frequency, bandwidth, and sampling rate of the RF signal, the resolution bandwidth set by the user, and the window function type.

[0009] The downsampled IQ signal is obtained by using a cascaded FIR filter based on the downsampling coefficient;

[0010] Based on the actual sampling rate, the downsampled IQ signal is divided into either a signal that is not oversampled or a signal that is oversampled:

[0011] If the signal is divided into signals that have not exceeded the sampling rate, the downsampled IQ signal is a single serial IQ signal. The serial IQ signal is then subjected to overlapping segmentation to obtain the serial segmented time-domain signal.

[0012] If the signal is divided into oversampling rate signals, then the downsampling IQ signals are parallel. One IQ signal, for parallel The IQ signals are subjected to overlapping segmentation to obtain parallel segmented time-domain signals, where, This indicates the multiple of the clock frequency corresponding to the actual sampling rate under oversampling;

[0013] Apply a window function of the actual FFT length to the serially segmented or parallelly segmented time-domain signal and perform FFT transformation to obtain multiple sets of frequency-domain signals. Calculate the amplitude of the multiple sets of frequency-domain signals and perform averaging to obtain the frequency-synthesized signal.

[0014] The frequency synthesized signal is logarithmically transformed to obtain a logarithmically processed frequency domain signal, and video filtering is performed based on the logarithmically processed frequency domain signal to obtain a real-time spectrum signal.

[0015] Furthermore, the actual sampling rate is expressed as:

[0016] ;

[0017] In the formula, Indicates the actual sampling rate. Indicates the bandwidth of the radio frequency signal. , Indicates the starting frequency of the radio frequency signal. Indicates the termination frequency of the radio frequency signal. This indicates the upper limit of the sampling rate supported by the real-time spectrum. Indicates the upper limit of resolution bandwidth supported by the real-time spectrum;

[0018] The actual FFT length is expressed as:

[0019] ;

[0020] In the formula, Indicates the actual FFT length. This indicates the resolution bandwidth set by the user. This represents the main lobe width of the window function, where the main lobe width is determined by the window function type.

[0021] The downsampling coefficient is expressed as:

[0022] ;

[0023] In the formula, Indicates the downsampling coefficient. This indicates the sampling rate of the radio frequency signal.

[0024] Furthermore, the cascaded FIR filter includes L downsampling filters and an L+1-to-1 data selector, wherein,

[0025] The first downsampling filter is used to implement... Data extraction multiple times;

[0026] The 2nd to Lth downsampling filters are all used to achieve a 2x data extraction.

[0027] The L+1 to 1 data selector is used to select the output of each stage of the cascaded FIR filter based on the downsampling coefficient, determine the effective output port of the downsampled signal, and realize configurable gating of the output data of different filter stages.

[0028] Furthermore, the cascaded FIR filter supports a downsampling coefficient of 1 or 1. ,in This indicates the number of downsampling filters that use 2x data extraction;

[0029] The selection rule for the L+1 to 1 data selector is as follows:

[0030] When the downsampling coefficient is 1, the downsampled IQ signal is output directly;

[0031] When the sampling coefficient is hour:

[0032] If n=0, it means that the data extraction multiple includes one. The factor will output the radio frequency signal from the first downsampling filter;

[0033] If n≠0, it means that the data decimation factor contains n factors of 2, and the RF signal is output from the (n+1)th downsampling filter.

[0034] Furthermore, based on the actual sampling rate, the downsampled IQ signal is divided into signals that are not oversampled or signals that are oversampled, including:

[0035] If the actual downsampling coefficient is 1, then the downsampled IQ signal will be divided into an oversampling rate signal.

[0036] If the actual downsampling factor is Then the downsampled IQ signal will be divided into signals that have not exceeded the sampling rate.

[0037] Further, the step of performing overlapping segmentation on a single serial IQ signal to obtain a serially segmented time-domain signal includes:

[0038] Calculate the maximum overlap rate of the downsampled IQ signals;

[0039] Based on the maximum overlap rate, one IQ signal is overlapped and divided into two IQ data streams, which are used as the serially segmented time-domain signals.

[0040] The parallel The IQ signals are overlapped and segmented to obtain parallel segmented time-domain signals, including:

[0041] Calculate the maximum overlap rate of the downsampled IQ signals;

[0042] Based on the maximum overlap rate The IQ signals are overlapped and divided into 2 Parallel IQ data is obtained by converting parallel data to serial data and then converting it to 2. The serial IQ data is used as a time-domain signal after parallel partitioning.

[0043] Furthermore, a window function of the actual FFT length is applied to the serially segmented or parallelly segmented time-domain signal, and FFT transformation is performed to obtain multiple sets of frequency-domain signals. The amplitudes of the multiple sets of frequency-domain signals are calculated and averaged to obtain the frequency-synthesized signal, including:

[0044] A window function of the actual FFT length is applied to the serially segmented or parallelly segmented time-domain signal and then fed into the FFT module.

[0045] The FFT module matches the corresponding FFT parallel number based on the number of paths of the serially or parallelly partitioned time-domain signal, and uses... To truncate the frequency domain data output by the FFT module, an intermediate data truncation operation is performed, retaining data with indices 0 to 1. , ~ Multiple sets of frequency domain signals are obtained from the frequency domain data of -1;

[0046] The amplitude of multiple frequency domain signals is calculated using a coordinate rotation digital calculation method.

[0047] The multiple frequency domain signals are divided into several frames of data based on their amplitudes.

[0048] Determine the segmentation method of the time-domain signal:

[0049] If it is a time-domain signal after serial segmentation, then each frame of data is divided into two data blocks based on several frames of data. All data blocks are summed and the average of the summation results is stored in the first dual-port random access memory.

[0050] If the signal is a time-domain signal after parallel partitioning, then each frame of data is divided into 2 based on several frames of data. One data block;

[0051] Summing all data blocks and multiplying the sum by Then move to the left The average value of the bits is stored in the first dual-port random access memory, where... for The bit width after converting to decimal;

[0052] All frame data are sequentially accumulated and stored in the second dual-port random access memory;

[0053] Based on the second dual-port random access memory, the average data of all frames is processed according to the total number of frames to obtain the frequency synthesized signal.

[0054] Furthermore, the frequency-synthesized signal is logarithmically transformed using a piecewise lookup table method to obtain the logarithmically processed frequency domain signal, including:

[0055] The simulation data of the fixed-point logarithm of the frequency synthesized signal bit width is stored in a segmented lookup table. The address of the corresponding logarithmic data of the frequency synthesized signal is determined according to the segmented lookup table, and logarithmic transformation is performed to obtain the frequency domain signal after logarithmic processing.

[0056] The frequency domain signal after logarithmic processing is represented as follows:

[0057] ;

[0058] In the formula, The frequency domain signal after logarithmic processing. It is a logarithmic function with base 10. The amplitude of the frequency-synthesized signal. for The starting value of the segment interval in the segment lookup table. for The starting address of the segment interval in the segment lookup table. The conversion ratio is logarithmic.

[0059] Furthermore, a real-time spectrum signal is obtained by performing video filtering on the logarithmically processed frequency domain signal, including:

[0060] Set the cutoff frequency of the FIR filter according to the video bandwidth:

[0061] when When setting the cutoff frequency of the FIR filter. ;

[0062] when When setting the cutoff frequency of the FIR filter. ;

[0063] in, Indicates video bandwidth. , This indicates the number of FIR filter coefficients used for video filtering;

[0064] Based on the cutoff frequency of the FIR filter, the logarithmically processed frequency domain signal is stored in a FIFO (First In First Out). Data is retrieved sequentially from the FIFO. When the last data is retrieved, a counter is started, and zero-padding is applied to the FIFO output data. The counter's counting range is 0~ ;

[0065] After the zero-padding operation is completed, the zero-padding FIFO output data is input into the FIR filter used for video filtering, outputting real-time spectrum data and skipping the real-time spectrum data. One data point;

[0066] In a second aspect, the present invention provides a parameter-adaptive real-time spectrum analysis system for implementing the parameter-adaptive real-time spectrum analysis method described in the first aspect, comprising:

[0067] The signal acquisition module is used to acquire the radio frequency signal to be analyzed.

[0068] The parameter calculation module is used to calculate the actual sampling rate, actual FFT length, and downsampling coefficient based on the start frequency, end frequency, bandwidth, and sampling rate of the RF signal, the resolution bandwidth set by the user, and the window function type.

[0069] The downsampling filtering module is used to perform downsampling filtering based on the downsampling coefficients using cascaded FIR filters to obtain the downsampled IQ signal;

[0070] The time-domain segmentation module is used to divide the downsampled IQ signal into either a non-oversampled-rate signal or an oversampled-rate signal based on the actual sampling rate. If it is divided into a non-oversampled-rate signal, the downsampled IQ signal is a single serial IQ signal, which is then subjected to overlap segmentation to obtain the serially segmented time-domain signal. If it is divided into an oversampled-rate signal, the downsampled IQ signal is parallel. One IQ signal, for parallel The IQ signals are subjected to overlapping segmentation to obtain parallel segmented time-domain signals, where, This indicates the multiple of the clock frequency corresponding to the actual sampling rate under oversampling;

[0071] The windowed fast Fourier transform and frequency synthesis module is used to apply a window function of the actual FFT length to the serially segmented or parallelly segmented time-domain signal and perform FFT transformation to obtain multiple sets of frequency-domain signals. The amplitude of the multiple sets of frequency-domain signals is calculated and averaged to obtain the frequency-synthesized signal.

[0072] The logarithmic transformation and video filtering module is used to perform logarithmic transformation on the frequency synthesized signal to obtain a logarithmically processed frequency domain signal, and to perform video filtering on the logarithmically processed frequency domain signal to obtain a real-time spectrum signal.

[0073] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0074] This invention dynamically calculates the actual sampling rate based on the bandwidth of the radio frequency signal, the user-defined resolution bandwidth, and the window function type. Based on the actual sampling rate, the downsampled IQ signal is divided into undersampled or oversampled signals for time-domain segmentation. A unified core of 2M serial FFT processing is used for time-frequency conversion, simplifying the hardware implementation complexity for different actual sampling rates. Compared to multi-channel parallel 2D FFT algorithm hardware architectures, this reduces data processing time and resource overhead. The oversampling path utilizes a parallel data processing method with M channels to overcome clock frequency limitations. Finally, through averaging multiple sets of frequency domain signals and video filtering, it meets the real-time spectrum analysis needs of users with high dynamic range, solving the problems of small bandwidth, long data processing time, and complex structure in existing real-time spectrum analysis technologies.

[0075] Logarithmic transformation using segmented lookup tables can further reduce resource overhead while maintaining accuracy. Attached Figure Description

[0076] Figure 1 This is a flowchart illustrating a parameter-adaptive real-time spectrum analysis method provided in an embodiment of the present invention;

[0077] Figure 2This is a schematic diagram of the downsampling filtering principle provided in an embodiment of the present invention.

[0078] Figure 3 This is a schematic diagram of the time-domain segmentation principle provided in an embodiment of the present invention.

[0079] Figure 4 This is a schematic diagram of a serial time-domain segmented data stream provided in an embodiment of the present invention.

[0080] Figure 5 This is a schematic diagram of the serial-to-parallel conversion principle provided in an embodiment of the present invention.

[0081] Figure 6 This is a schematic diagram of a parallel time-domain segmented data stream provided in an embodiment of the present invention.

[0082] Figure 7 This is a schematic diagram of the windowing principle provided in an embodiment of the present invention.

[0083] Figure 8 This is a schematic diagram of the FFT conversion principle provided in the embodiment of the present invention.

[0084] Figure 9 This is a schematic diagram illustrating the principle of frequency synthesis provided in an embodiment of the present invention.

[0085] Figure 10 This is a schematic diagram illustrating the principle of logarithmic transformation provided in an embodiment of the present invention.

[0086] Figure 11 This is a schematic diagram of the video filtering principle provided in an embodiment of the present invention. Detailed Implementation

[0087] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features therein are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. Where there is no conflict, the embodiments and technical features therein can be combined with each other. Additionally, the character " / " generally indicates that the preceding and following objects have an "or" relationship.

[0088] Example 1

[0089] like Figure 1 As shown in the figure, this embodiment introduces a parameter-adaptive real-time spectrum analysis method, including:

[0090] Step 1: Calculate the actual sampling rate, actual FFT length, and downsampling coefficient based on the start frequency, end frequency, bandwidth, and sampling rate of the RF signal, the resolution bandwidth set by the user, and the window function type.

[0091] This invention dynamically calculates the actual sampling rate, actual FFT length, and downsampling coefficient based on the input RF signal and feeds them back to the FPGA, realizing adaptive configuration of system parameters. This allows subsequent processing resources to be precisely matched with the current signal characteristics, ensuring performance while avoiding resource fixation and waste, and improving the system's flexibility and efficiency.

[0092] Step 2: Based on the downsampling coefficients, use cascaded FIR filters to perform downsampling filtering to obtain the downsampled IQ signal.

[0093] This invention utilizes cascaded FIR filters based on downsampling coefficients in FPGAs for downsampling filtering, which can efficiently achieve multi-level downsampling. While effectively suppressing aliasing interference and maintaining the signal waveform, it significantly reduces the rate and load of subsequent data processing links, laying the foundation for real-time processing of high sampling rate signals.

[0094] Step 3: Based on the actual sampling rate, divide the downsampled IQ signal into signals with or without oversampling rate:

[0095] If the signal is divided into signals that have not exceeded the sampling rate, the downsampled IQ signal is a single serial IQ signal. The serial IQ signal is then subjected to overlapping segmentation to obtain the serial segmented time-domain signal.

[0096] If the signal is divided into oversampling rate signals, then the downsampling IQ signals are parallel. One IQ signal, for parallel The IQ signals are subjected to overlapping segmentation to obtain parallel segmented time-domain signals, where, This indicates the multiple of the clock frequency corresponding to the actual sampling rate under oversampling.

[0097] This invention adaptively divides the signal path into signals with or without oversampling rate based on the actual sampling rate, constructing a scalable parallel processing architecture. This architecture enables the system to efficiently process signals at normal rates and improve data throughput by M times through parallel processing paths, thus successfully breaking through the fundamental limitation of traditional single-channel serial processing on real-time analysis bandwidth.

[0098] Step 4: Apply a window function of the actual FFT length to the serially segmented or parallelly segmented time-domain signal and perform FFT transformation to obtain multiple sets of frequency-domain signals. Calculate the amplitude of the multiple sets of frequency-domain signals and perform averaging to obtain the frequency-synthesized signal.

[0099] This invention applies a window function to the segmented time-domain signal and performs FFT transformation. By calculating the amplitude of multiple sets of frequency-domain signals and averaging them, it effectively suppresses noise and random interference, improves the signal-to-noise ratio and stability of spectrum measurement, and simplifies the hardware implementation complexity for signals from different paths through a unified FFT processing core.

[0100] Step 5: Perform a logarithmic transformation on the frequency synthesized signal to obtain a logarithmically processed frequency domain signal, and perform video filtering based on the logarithmically processed frequency domain signal to obtain a real-time spectrum signal.

[0101] This invention utilizes a segmented lookup table method for logarithmic transformation. Compared to direct calculation, this method can complete nonlinear operations in FPGA with extremely low logic resource overhead and extremely high speed, significantly improving processing efficiency. Furthermore, by combining video filtering to smooth the frequency domain signal after logarithmic processing, a high-quality real-time spectrum signal that is more conducive to observation and analysis is finally output.

[0102] Example 2

[0103] Based on the same inventive concept as Embodiment 1, this embodiment describes the implementation steps of a parameter-adaptive real-time spectrum analysis method, including:

[0104] Step 1: Calculate the actual sampling rate, actual FFT length, and downsampling coefficient based on the start frequency, end frequency, bandwidth, and sampling rate of the RF signal, the resolution bandwidth set by the user, and the window function type.

[0105] In this embodiment, the specific values ​​of the input parameters are input to the ARM terminal to calculate the actual sampling rate, actual FFT length, and downsampling coefficient, and then fed back to the FPGA. The input parameters include the start frequency and end frequency of the radio frequency signal, the type of window function, and the resolution bandwidth set by the user.

[0106] In this embodiment, the actual sampling rate is expressed as:

[0107] ;

[0108] In the formula, Indicates the actual sampling rate. Indicates the bandwidth of the radio frequency signal. , Indicates the starting frequency of the radio frequency signal. Indicates the termination frequency of the radio frequency signal. This indicates the upper limit of the sampling rate supported by the real-time spectrum. This indicates the upper limit of the resolution bandwidth supported by the real-time spectrum. In this embodiment, the upper limit of the sampling rate supported by the real-time spectrum is 1.2288 Gsps, and the upper limit of the resolution bandwidth supported by the real-time spectrum is 1 GHz.

[0109] In this embodiment, the actual FFT length is represented as:

[0110] ;

[0111] In the formula, Indicates the actual FFT length. This indicates the resolution bandwidth set by the user. This represents the main lobe width of the window function, where the main lobe width is determined by the window function type.

[0112] In this embodiment, the downsampling coefficient is expressed as:

[0113] ;

[0114] In the formula, Indicates the downsampling coefficient. This indicates the sampling rate of the radio frequency signal.

[0115] Step 2: Based on the downsampling coefficients, use cascaded FIR filters to perform downsampling filtering to obtain the downsampled IQ signal.

[0116] The cascaded FIR filter includes L downsampling filters and an L+1 to 1 data selector, wherein the first downsampling filter is used to implement... The data is decimated by 2 times; the 2nd to Lth downsampling filters are all used to achieve 2x data decimation; the L+1 to 1 data selector is used to select the output of each stage of the cascaded FIR filter according to the downsampling coefficient, determine the effective output port of the downsampled signal, and realize the configurable selection of the output data of different filter stages.

[0117] The cascaded FIR filters support downsampling coefficients of 1 or 1. ,in This indicates the number of downsampling filters using 2x data decimation; the selection rule for the L+1 to 1 data selector is as follows:

[0118] When the downsampling coefficient is 1, the downsampled IQ signal is output directly;

[0119] When the sampling coefficient is hour:

[0120] If n=0, it means that the data extraction multiple includes one value. The factor determines the output of the radio frequency signal from the first downsampling filter;

[0121] If n≠0, it means that the data decimation factor contains n factors with a value of 2, and the radio frequency signal is output from the (n+1)th downsampling filter.

[0122] In this embodiment, The value is 5. The cascaded FIR filter includes 16 downsampling filters and a 17-to-1 data selector, wherein the first downsampling filter is used to implement... The downsampling filters are all used to achieve 2x data decimation; the 2nd to 16th downsampling filters are all used to achieve 2x data decimation; the 17-to-1 data selector is used to select the output of each stage of the cascaded FIR filter according to the downsampling coefficient, determine the effective output port of the downsampled signal, and realize configurable selection of the output data of different filter stages.

[0123] In this embodiment, the cascaded FIR filter supports a downsampling coefficient of 1 or 1. ,in This indicates the number of downsampling filters used with a 2x data decimation. The supported range is 0~15; the selection rule for the 17-to-1 data selector is as follows:

[0124] When the downsampling coefficient is 1, the downsampled IQ signal is output directly;

[0125] When the sampling coefficient is When n=0, it means that the data decimation factor includes a factor of 5, and the radio frequency signal is output from the first downsampling filter;

[0126] If n≠0, it means that the data decimation factor contains n factors with a value of 2, and the radio frequency signal is output from the (n+1)th downsampling filter.

[0127] The block diagram of the downsampling filter provided in this embodiment is as follows: Figure 2 As shown, the output rates include 17 types: 1.2288 GSps, 245.76 MSps, 122.88 MSps, 61.44 MSps, 30.72 MSps, 15.36 MSps, 7.68 MSps, 3.84 MSps, 1.92 MSps, 960 KSps, 480 KSps, 240 KSps, 120 KSps, 60 KSps, 30 KSps, 15 KSps, and 7.5 KSps. When the data output rate is 245.76 MSps or lower, the data channel is represented as X1; when the data output rate is 1228.8 MSps, the data channel is represented as X5. Figure 2In the diagram, FIR_5 is a 5x downsampled FIR filter with 401 filter coefficients, implemented using an IP core. FIR_2_1 through FIR_2_15 are 2x downsampled FIR filters, each with 201 filter coefficients. However, because the input data rate of each filter decreases by a factor of 2, the IP core generation methods differ for each filter. Compared to FIR_1, which uses 51 DSP resources, FIR_2_15 uses only 1 DSP resource.

[0128] Step 3: Divide the downsampled IQ signal into signals with or without oversampling rate, based on the actual sampling rate:

[0129] If the signal is divided into signals that have not exceeded the sampling rate, the downsampled IQ signal is a single serial IQ signal. The serial IQ signal is then subjected to overlapping segmentation to obtain the serial segmented time-domain signal.

[0130] If the signal is divided into oversampling rate signals, then the downsampling IQ signals are parallel. One IQ signal, for parallel The IQ signals are subjected to overlapping segmentation to obtain parallel segmented time-domain signals, where, This indicates the multiple of the clock frequency corresponding to the actual sampling rate under oversampling.

[0131] In this embodiment, if the actual downsampling coefficient is 1, the downsampled IQ signal is divided into an oversampling rate signal; if the actual downsampling coefficient is... Then the downsampled IQ signal will be divided into signals that have not exceeded the sampling rate.

[0132] In this embodiment, if the signal is divided into a non-oversampling rate signal, the downsampled IQ signal is a single serial IQ signal. Overlapping segmentation is performed on this single serial IQ signal to obtain the serially segmented time-domain signal, including:

[0133] The maximum overlap rate of the radio frequency signals is calculated by the host computer.

[0134] One IQ signal is divided into two IQ data streams based on the maximum overlap rate, which are then used as the serially segmented time-domain signal.

[0135] The overall temporal domain segmentation principle block diagram provided in this embodiment is as follows: Figure 3 As shown, the downsampled signal is processed according to the actual FFT length. and maximum overlap rate The data is divided into blocks. Data block 1 is output from the first path, data block 2 from the second path, data block 3 from the first path, data block 4 from the second path, and so on. Assuming the actual FFT length is 1024 and the overlap rate is 50%, ... Figure 4This is a schematic diagram of serial time-domain partitioning. Figure 4 The serial segmentation control is mainly responsible for dividing a continuous time-domain data stream into segments according to the given actual FFT length. and maximum overlap rate Divide into several actual FFT lengths The data blocks are processed, and a specified overlap is maintained between adjacent windows. After serial splitting is initiated, the first data path is output and counter 1 is started. When counter 1 reaches the actual FFT length... When the first data block ends, counter 1 is reset to 0; simultaneously, when counter 1 is... When the second data channel is output, counter 2 is started. When counter 2 reaches the actual FFT length... When the second data block ends, counter 2 is reset to 0, and this cycle repeats.

[0136] In this embodiment, if the signal is divided into oversampling rate signals, the downsampled IQ signals are 5 parallel IQ signals. Overlapping segmentation is performed on these 5 parallel IQ signals to obtain the parallel segmented time-domain signal, including:

[0137] The maximum overlap rate of the radio frequency signals is calculated by the host computer.

[0138] Based on the maximum overlap rate The IQ signal is overlapped and split into 10 parallel IQ data streams, which are then converted into 10 serial IQ data streams as the time-domain signals after parallel splitting. The overlap logic for data splitting is the same as that for serial splitting. The difference is that after parallel splitting, a FIFO buffer with a depth of 4096 is added after each parallel IQ data stream. The system waits until all 5 parallel IQ data streams are split before performing serial-to-parallel conversion. The specific serial-to-parallel conversion principle diagram is shown below. Figure 5 As shown. =1024, Taking 50% as an example, parallel data stream partitioning is as follows: Figure 6 As shown, D0 represents the 0th data point of the downsampled IQ signal, D1 represents the 1st data point of the downsampled IQ signal, and so on.

[0139] Step 4: Apply a window function of the actual FFT length to the serially segmented or parallelly segmented time-domain signal and perform FFT transformation to obtain multiple sets of frequency-domain signals. Calculate the amplitude of the multiple sets of frequency-domain signals and perform averaging to obtain the frequency synthesized signal.

[0140] Step 4.1: Apply a window function of the actual FFT length to the serially segmented or parallelly segmented time-domain signal and send it to the FFT module.

[0141] like Figure 7This is a block diagram illustrating the principle of windowing. The specific method for applying the window function is to configure two multipliers for each data path, respectively multiplying the I-channel signal and the Q-channel signal. The window functions of different lengths are multiplied accordingly, and the result is output. Window parameters are stored in dual-port random access memory (DPRAM) and the read window parameters are controlled according to the input time of the segmented time-domain signal and the selection of the window function, where raddr is the read window function address and rdata is the read window function parameter. The following table shows the phase parameters of the window function. The wider the main lobe, the wider the effective noise bandwidth, but the worse the frequency resolution is under the same frequency resolution; however, the wider the main lobe, the higher the amplitude recognition accuracy and the better the power accuracy. The real-time spectrum supports various parameters of different window functions, as shown in Table 1.

[0142] Table 1. Parameters of different window functions supported by real-time spectrum

[0143]

[0144] Step 4.2: Use the FFT module to match the corresponding FFT parallel number based on the number of paths of the serially segmented or parallelly segmented time-domain signal, and use... To truncate the frequency domain data output by the FFT module, an intermediate data truncation operation is performed, retaining data with indices 0 to 1. , ~ Multiple sets of frequency domain signals are obtained from the frequency domain data of -1.

[0145] like Figure 8 The diagram shown illustrates the principle of FFT transformation. Due to resource limitations, the number of points supported per channel is limited to a certain range. n is 2~14. In this embodiment, the data retention ratio is 0.81. The partial FFT length and the length of the retained frequency domain data are shown in Table 2.

[0146] Table 2. Partial FFT lengths and lengths of retained frequency domain data

[0147]

[0148] Step 4.3: Calculate the amplitude of multiple frequency domain signals using the coordinate rotation digital calculation method.

[0149] Step 4.4: Divide the multiple frequency domain signals into several frames of data based on the amplitude of the multiple frequency domain signals.

[0150] Step 4.5: Determine the segmentation method of the time-domain signal:

[0151] If it is a time-domain signal after serial segmentation, then each frame of data is divided into two data blocks based on several frames of data. All data blocks are summed and the average of the summation results is stored in the first dual-port random access memory.

[0152] If the signal is a time-domain signal after parallel partitioning, then each frame of data is divided into 2 based on several frames of data. One data block,

[0153] Step 4.6: Summate all data blocks and multiply the sum by... Then move to the left The average value of the bits is stored in the second dual-port random access memory, where... for The bit width after conversion to decimal.

[0154] Step 4.7: Accumulate all frame data sequentially and store them into the second dual-port random access memory.

[0155] Step 4.8: Based on the first dual-port random access memory and the second dual-port random access memory, average all frame data according to the total number of frames to obtain the frequency synthesized signal.

[0156] Figure 9 This is a schematic diagram illustrating the principle of frequency synthesis provided in an embodiment of the present invention.

[0157] Step 5: Perform logarithmic transformation on the frequency synthesized signal to obtain the logarithmically processed frequency domain signal, and perform video filtering based on the logarithmically processed frequency domain signal to obtain the real-time spectrum signal.

[0158] In this embodiment, the frequency-synthesized signal is logarithmically transformed using a piecewise lookup table method to obtain a logarithmically processed frequency domain signal, including:

[0159] The simulation data of the fixed-point logarithm of the frequency synthesized signal bit width is stored in a segmented lookup table. The address of the corresponding logarithmic data of the frequency synthesized signal is determined according to the segmented lookup table, and logarithmic transformation is performed to obtain the frequency domain signal after logarithmic processing.

[0160] The frequency domain signal after logarithmic processing is represented as follows:

[0161] ;

[0162] In the formula, The frequency domain signal after logarithmic processing. It is a logarithmic function with base 10. The amplitude of the frequency-synthesized signal. for The starting value of the segment interval in the segment lookup table. for The starting address of the segment interval in the segment lookup table. The conversion ratio is logarithmic.

[0163] In this embodiment, the frequency-synthesized signal The bit width is 18 bits, representing a range of 0 to 262143. This range is divided into four intervals, each with a corresponding logarithmic depth. For example, [0, 256) corresponds to 256 logarithmic data points, which is the conversion ratio. The value is 1, and [256, 2048) corresponds to 256 logarithmic data points, meaning the conversion ratio is 7, which means 7 input data points correspond to 1 logarithm. Starting address To find the starting address of different intervals in the lookup table index, the lookup table addressing mode is start address. Add range address The intervals and their corresponding parameters are shown in Table 3.

[0164] Table 3. Each interval and its corresponding parameters

[0165]

[0166] The logarithmic lookup table calculation was performed offline using MATLAB, saved in COE file format, and then loaded into read-only memory (ROM). Determine the range of the input interval and obtain the result. and ,Will Store in a FIFO, and use a divider to calculate the range address. ,in This represents the starting value of the input data within its specified interval. Adding the two inputs together yields the address of the logarithmic lookup table for the input data. The final result is the data in decibel form. , .like Figure 10 This is a schematic diagram of piecewise logarithmic transformation. The input data width is 18 bits, and the output data width is 16 bits.

[0167] In this embodiment, a real-time spectrum signal is obtained by video filtering based on the logarithmically processed frequency domain signal. The video filtering principle diagram is shown below. Figure 11 As shown, it includes:

[0168] The cutoff frequency of the FIR filter is set according to the video bandwidth. The configuration of the FIR filter used for video filtering consists of 201 filter coefficients, which are loaded via an ARM. When setting the filter cutoff frequency ;when When setting the filter cutoff frequency .

[0169] Based on the cutoff frequency of the FIR filter, the logarithmically processed frequency domain signal is stored in a FIFO. Data is retrieved sequentially from the FIFO. When the last data is retrieved, a counter is started, and zero-padding is applied to the FIFO output data. The counter's counting range is 0 to 99.

[0170] After the zero-padding operation is completed, the zero-padding FIFO output data is input into the FIR filter used for video filtering, and real-time spectrum data is output.

[0171] In this embodiment, the FIR filter output data of the video filtering is 40 bits wide, which needs to be truncated to form 16 bits of data, and the first 100 data points of the real-time spectrum data are skipped.

[0172] Example 3

[0173] Based on the same inventive concept as Embodiment 1, this embodiment introduces a parameter-adaptive real-time spectrum analysis system for implementing the parameter-adaptive real-time spectrum analysis method described in Embodiment 1 or 2, including:

[0174] The parameter calculation module is used to calculate the actual sampling rate, actual FFT length, and downsampling coefficient based on the start frequency, end frequency, bandwidth, and sampling rate of the RF signal, the resolution bandwidth set by the user, and the window function type.

[0175] The downsampling filtering module is used to perform downsampling filtering based on the downsampling coefficients using cascaded FIR filters to obtain the downsampled IQ signal;

[0176] The time-domain segmentation module is used to divide the downsampled IQ signal into either a non-oversampled-rate signal or an oversampled-rate signal based on the actual sampling rate. If it is divided into a non-oversampled-rate signal, the downsampled IQ signal is a single serial IQ signal, which is then subjected to overlap segmentation to obtain the serially segmented time-domain signal. If it is divided into an oversampled-rate signal, the downsampled IQ signal is parallel. One IQ signal, for parallel The IQ signals are subjected to overlapping segmentation to obtain parallel segmented time-domain signals, where, This indicates the multiple of the clock frequency corresponding to the actual sampling rate under oversampling;

[0177] The windowed fast Fourier transform and frequency synthesis module is used to apply a window function of the actual FFT length to the serially segmented or parallelly segmented time-domain signal and perform FFT transformation to obtain multiple sets of frequency-domain signals. The amplitude of the multiple sets of frequency-domain signals is calculated and averaged to obtain the frequency-synthesized signal.

[0178] The logarithmic transformation and video filtering module is used to perform logarithmic transformation on the frequency synthesized signal to obtain a logarithmically processed frequency domain signal, and to perform video filtering on the logarithmically processed frequency domain signal to obtain a real-time spectrum signal.

[0179] The specific functions of each module described above are explained in the relevant content of the methods in Embodiment 1 or 2, and will not be repeated here.

[0180] Example 4

[0181] Based on the same inventive concept as other embodiments, this embodiment describes a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the steps of the methods of Embodiment 1 or 2 described above.

[0182] Example 5

[0183] Based on the same inventive concept as other embodiments, this embodiment introduces a computer program product, including computer instructions that, when executed by a processor, implement the steps of the methods described in Embodiment 1 or 2 above.

[0184] In summary, this invention divides the downsampled IQ signal into non-oversampled or oversampled signals based on the actual sampling rate for time-domain segmentation. It utilizes 2M serial FFTs for time-frequency conversion and a segmented lookup table for logarithmic transformation, thereby reducing resource overhead while maintaining real-time performance. Specifically, the non-oversampled path uses single-channel overlapping segmentation of serial IQ signals to reduce computational complexity, while the oversampled path utilizes parallel M-channel IQ signal processing to overcome clock frequency limitations. Finally, through averaging of multiple sets of frequency domain signals and video filtering, it meets the requirements for high dynamic range real-time spectrum analysis, solving the problems of small bandwidth, long data processing time, and complex structure in existing real-time spectrum analysis technologies.

[0185] This invention achieves adaptive sampling rate adjustment and optimized hardware resource configuration by dynamically dividing the non-oversampled rate signal or the oversampled rate signal and matching it with a cascaded FIR filter downsampling structure. The non-oversampled path uses serial single-channel IQ signal overlap segmentation to reduce FFT computation complexity, while the oversampled path utilizes parallel M-channel IQ signal processing to overcome clock frequency limitations, and is further enhanced by a downsampling coefficient of 1 or... The cascaded filter selection rules ensure that high-precision spectrum analysis can be maintained under different downsampling requirements while reducing FPGA resource consumption. At the same time, the unified FFT processing core simplifies the hardware implementation complexity for signals from different paths.

[0186] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0187] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0188] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0189] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0190] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A parameter-adaptive real-time spectrum analysis method, characterized in that, include: Acquire the radio frequency signal to be analyzed; The actual sampling rate, actual FFT length, and downsampling coefficient are calculated based on the start frequency, end frequency, bandwidth, and sampling rate of the RF signal, the resolution bandwidth set by the user, and the window function type. The downsampled IQ signal is obtained by using a cascaded FIR filter based on the downsampling coefficient; Based on the actual sampling rate, the downsampled IQ signal is divided into either a signal that is not oversampled or a signal that is oversampled: If the signal is divided into signals that have not exceeded the sampling rate, the downsampled IQ signal is a single serial IQ signal. The serial IQ signal is then subjected to overlapping segmentation to obtain the serial segmented time-domain signal. If the signal is divided into oversampling rate signals, then the downsampling IQ signals are parallel. One IQ signal, for parallel The IQ signals are subjected to overlapping segmentation to obtain parallel segmented time-domain signals, where, This indicates the multiple of the clock frequency corresponding to the actual sampling rate under oversampling; Apply a window function of the actual FFT length to the serially segmented or parallelly segmented time-domain signal and perform FFT transformation to obtain multiple sets of frequency-domain signals. Calculate the amplitude of the multiple sets of frequency-domain signals and perform averaging to obtain the frequency-synthesized signal. The frequency synthesized signal is logarithmically transformed to obtain a logarithmically processed frequency domain signal, and video filtering is performed based on the logarithmically processed frequency domain signal to obtain a real-time spectrum signal.

2. The parameter-adaptive real-time spectrum analysis method according to claim 1, characterized in that, The actual sampling rate is expressed as: ； In the formula, Indicates the actual sampling rate. Indicates the bandwidth of the radio frequency signal. , Indicates the starting frequency of the radio frequency signal. Indicates the termination frequency of the radio frequency signal. This indicates the upper limit of resolution bandwidth supported by the real-time spectrum. Indicates the upper limit of the sampling rate supported by the real-time spectrum; The actual FFT length is expressed as: ； In the formula, Indicates the actual FFT length. This indicates the resolution bandwidth set by the user. This represents the main lobe width of the window function, where the main lobe width is determined by the window function type. The downsampling coefficient is expressed as: ； In the formula, Indicates the downsampling coefficient. This indicates the sampling rate of the radio frequency signal.

3. The parameter-adaptive real-time spectrum analysis method according to claim 2, characterized in that, The cascaded FIR filter includes L downsampling filters and L+1 to 1 data selectors, wherein, The first downsampling filter is used to implement... Data extraction multiple times; The 2nd to Lth downsampling filters are all used to achieve a 2x data extraction. The L+1 to 1 data selector is used to select the output of each stage of the cascaded FIR filter based on the downsampling coefficient, determine the effective output port of the downsampled signal, and realize configurable gating of the output data of different filter stages.

4. The parameter-adaptive real-time spectrum analysis method according to claim 3, characterized in that, The cascaded FIR filters support downsampling coefficients of 1 or 1. ,in This indicates the number of downsampling filters that use 2x data extraction; The selection rule for the L+1 to 1 data selector is as follows: When the downsampling coefficient is 1, the downsampled IQ signal is output directly; When the sampling coefficient is hour: If n=0, it means that the data extraction multiple includes one value. The factor determines the output of the radio frequency signal from the first downsampling filter; If n≠0, it means that the data decimation factor contains n factors with a value of 2, and the radio frequency signal is output from the (n+1)th downsampling filter.

5. The parameter-adaptive real-time spectrum analysis method according to claim 4, characterized in that, Based on the actual sampling rate, the downsampled IQ signal is divided into signals that are not oversampled or signals that are oversampled, including: If the actual downsampling coefficient is 1, then the downsampled IQ signal will be divided into an oversampling rate signal. If the actual downsampling factor is Then the downsampled IQ signal will be divided into signals that have not exceeded the sampling rate.

6. The parameter-adaptive real-time spectrum analysis method according to claim 5, characterized in that, The process of performing overlapping segmentation on a single serial IQ signal to obtain a serially segmented time-domain signal includes: Calculate the maximum overlap rate of the downsampled IQ signals; One IQ signal is divided into two IQ data streams based on the maximum overlap rate, which are then used as the serially segmented time-domain signal. The parallel The IQ signals are overlapped and segmented to obtain parallel segmented time-domain signals, including: Calculate the maximum overlap rate of the downsampled IQ signals; Based on the maximum overlap rate The IQ signals are overlapped and divided into 2 Parallel IQ data is obtained by converting parallel data to serial data and then converting it to 2. The serial IQ data is used as a time-domain signal after parallel partitioning.

7. The parameter-adaptive real-time spectrum analysis method according to claim 6, characterized in that, Applying a window function of the actual FFT length to the serially segmented or parallelly segmented time-domain signal and performing FFT transformation yields multiple sets of frequency-domain signals. The amplitudes of these multiple frequency-domain signals are calculated and averaged to obtain the frequency-synthesized signal, including: A window function of the actual FFT length is applied to the serially segmented or parallelly segmented time-domain signal and then fed into the FFT module. The FFT module matches the corresponding FFT parallel number based on the number of paths of the serially or parallelly partitioned time-domain signal, and uses... To truncate the frequency domain data output by the FFT module, an intermediate data truncation operation is performed, retaining data with indices 0 to 1. , ~ Multiple sets of frequency domain signals are obtained from the frequency domain data of -1; The amplitude of multiple frequency domain signals is calculated using a coordinate rotation digital calculation method. The multiple frequency domain signals are divided into several frames of data based on their amplitudes. Determine the segmentation method of the time-domain signal: If it is a time-domain signal after serial segmentation, then each frame of data is divided into two data blocks based on several frames of data. All data blocks are summed and the average of the summation results is stored in the first dual-port random access memory. If the signal is a time-domain signal after parallel partitioning, then each frame of data is divided into 2 based on several frames of data. One data block; Summing all data blocks and multiplying the sum by Then move to the left The average value of the bits is stored in the first dual-port random access memory, where... for The bit width after converting to decimal; All frame data are sequentially accumulated and stored in the second dual-port random access memory; Based on the second dual-port random access memory, the average data of all frames is processed according to the total number of frames to obtain the frequency synthesized signal.

8. The parameter-adaptive real-time spectrum analysis method according to claim 7, characterized in that, The frequency-synthesized signal is logarithmically transformed to obtain the logarithmically processed frequency domain signal, including: The simulation data of the fixed-point logarithm of the frequency synthesized signal bit width is stored in a segmented lookup table. The address of the corresponding logarithmic data of the frequency synthesized signal is determined according to the segmented lookup table, and logarithmic transformation is performed to obtain the frequency domain signal after logarithmic processing. The frequency domain signal after logarithmic processing is represented as follows: ； In the formula, The frequency domain signal after logarithmic processing. It is a logarithmic function with base 10. The amplitude of the frequency-synthesized signal. for The starting value of the segment interval in the segment lookup table. for The starting address of the segment interval in the segment lookup table. The conversion ratio is logarithmic.

9. The parameter-adaptive real-time spectrum analysis method according to claim 8, characterized in that, Real-time spectrum signals are obtained by video filtering of the frequency domain signal after logarithmic processing, including: Set the cutoff frequency of the FIR filter according to the video bandwidth: when When setting the cutoff frequency of the FIR filter. ; when When setting the cutoff frequency of the FIR filter. ; in, Indicates video bandwidth. , This indicates the number of FIR filter coefficients used for video filtering; Based on the cutoff frequency of the FIR filter, the logarithmically processed frequency domain signal is stored in a FIFO. Data is retrieved sequentially from the FIFO. When the last data is retrieved, a counter is started, and zero-padding is applied to the FIFO output data. The counter's counting range is 0~ ; After the zero-padding operation is completed, the zero-padding FIFO output data is input into the FIR filter used for video filtering, outputting real-time spectrum data and skipping the real-time spectrum data. Data points.

10. A parameter-adaptive real-time spectrum analysis system, characterized in that, The steps for implementing the parameter adaptive real-time spectrum analysis method according to any one of claims 1-9 include: The signal acquisition module is used to acquire the radio frequency signal to be analyzed. The parameter calculation module is used to calculate the actual sampling rate, actual FFT length, and downsampling coefficient based on the start frequency, end frequency, bandwidth, and sampling rate of the RF signal, the resolution bandwidth set by the user, and the window function type. The downsampling filtering module is used to perform downsampling filtering based on the downsampling coefficients using cascaded FIR filters to obtain the downsampled IQ signal; The time-domain segmentation module is used to divide the downsampled IQ signal into either a non-oversampled-rate signal or an oversampled-rate signal based on the actual sampling rate. If it is divided into a non-oversampled-rate signal, the downsampled IQ signal is a single serial IQ signal, which is then subjected to overlap segmentation to obtain the serially segmented time-domain signal. If it is divided into an oversampled-rate signal, the downsampled IQ signal is parallel. One IQ signal, for parallel The IQ signals are subjected to overlapping segmentation to obtain parallel segmented time-domain signals, where, This indicates the multiple of the clock frequency corresponding to the actual sampling rate under oversampling; The windowed fast Fourier transform and frequency synthesis module is used to apply a window function of the actual FFT length to the serially segmented or parallelly segmented time-domain signal and perform FFT transformation to obtain multiple sets of frequency-domain signals. The amplitude of the multiple sets of frequency-domain signals is calculated and averaged to obtain the frequency-synthesized signal. The logarithmic transformation and video filtering module is used to perform logarithmic transformation on the frequency synthesized signal to obtain a logarithmically processed frequency domain signal, and to perform video filtering on the logarithmically processed frequency domain signal to obtain a real-time spectrum signal.

Images