A phase demodulation method based on signal slicing and overlapping calculation of FFT and IFFT frequency domain filtering

Abstract

The present invention discloses a phase demodulation method based on signal slicing and overlapping calculation of FFT and IFFT frequency domain filtering, including zeroing unnecessary frequency components in the frequency domain after FFT to achieve efficient digital filtering; obtaining orthogonal signal pairs by frequency domain complex multiplication and IFFT; using data slicing, overlapping calculation and demodulation result reorganization to greatly suppress the "endpoint effect" existing in the FFT algorithm, thereby achieving phase demodulation. In the demodulation process of the present invention, unnecessary frequency components in the signal can be eliminated without traditional digital filtering, and orthogonal signal pairs can be obtained; the method of parallel calculation of multiple FFT / IFFT modules can greatly improve the efficiency of calculation, which is not only suitable for phase demodulation in offline conditions, but also can be used for real-time phase demodulation on digital processors such as FPGA; it is highly flexible, and different data slice lengths can be set for specific application backgrounds to accelerate the phase demodulation process.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to a phase demodulation method based on signal slicing and overlapping calculation of FFT and IFFT frequency domain filtering. Background Art

[0002] Phase demodulation is a signal processing technique used to recover the original phase information from a modulated signal. In phase modulation (PM) or frequency modulation (FM), information is encoded in the phase changes of the carrier signal. The purpose of phase demodulation is to detect and recover this information from the changing phase. Therefore, phase demodulation is essential in general situations involving the need to extract information from modulated signals, especially in systems where the information is encoded in the phase changes.

[0003] Phase demodulation is a basic signal processing technique that is mainly used in the following areas: Wireless communications: Phase demodulation is a common technique in wireless communications to recover information from modulated signals. Cell phones, wireless local area networks (WLAN), satellite communications, and radio broadcasting all use phase modulation and demodulation techniques. Radar systems: Radar uses phase modulation to measure the distance and speed of target objects. In synthetic aperture radar (SAR) and other advanced radar technologies, phase demodulation is used to process echo signals to obtain high-resolution images. Fiber-optic communications: Phase modulation and demodulation techniques are also very important in fiber-optic communications, especially in coherent optical communication systems, where they can increase data transmission rates and link distances. Modems: In traditional telephone line modems and modern broadband modems, phase modulation and demodulation techniques are used to encode and decode digital data in analog signals. Audio and video broadcasting: Frequency modulation in FM broadcasting essentially involves continuous changes in phase, so phase demodulation is a key step in recovering audio signals at the receiver. Navigation systems: Phase modulation techniques are used in the Global Positioning System (GPS) and other satellite navigation systems to provide precise positioning information. The receiver must be able to demodulate these signals to calculate the exact position. Scientific and Engineering Research: Phase demodulation is used for signal analysis in experimental devices in physics, astronomy, and other scientific fields, such as the application of interferometers in optics and quantum computing. Commonly used phase demodulation techniques usually include digital IQ demodulation, Hilbert transform phase demodulation, dual lock-in amplifier demodulation, etc.

[0004] In order to obtain better filtering effect for digital IQ demodulation, the frequency response of the digital low-pass filter must be designed to be very flat, which will consume a lot of digital processor hardware resources.

[0005] In Hilbert transform phase demodulation, the flatness requirement for the Hilbert transform filter is high, so relatively high computing resources are needed, especially in real-time or high-speed data processing.

[0006] In SPS phase demodulation, the requirements for the data acquisition system are very high, and the interval between adjacent channels must be very small.

[0007] In view of the above problems, the present invention provides a high-precision, high-parallelism phase demodulation method based on signal slicing and overlapping calculation of FFT and IFFT frequency domain filtering. Summary of the invention

[0008] The invention provides a high-precision and high-parallelism phase demodulation method of signal slicing and overlapping calculation based on FFT and IFFT frequency domain filtering.

[0009] The object of the present invention is to provide a phase demodulation method for signal slicing and overlapping calculation based on FFT and IFFT frequency domain filtering, the method comprising the following steps:

[0010] Step 1: Digitally sample the analog heterodyne interference signal. The sampling rate must meet the following requirements:

[0011] f s >2·BW

[0012] Where BW is the signal bandwidth and:

[0013] BW=2(2v p / λ+f v )

[0014] Step 2: Overlap and slice the sampled digital signal. The original data is a continuous one-dimensional digital signal. Slice the one-dimensional data into segments suitable for FFT. The data length must be 2 N , where N is a positive integer), and set the relevant parameters as follows: p1 = 2 N -(2 N / m+1);p2=p1–1;p3=p1;p4=2 N -p2 -1; p5 = (2 N -p1+1) / 2; p6=2 N +1-p5; where m is generally 2, 4, 8, or 16. When m=4 and N=8, the first data fragment is 1 to 256 of the original data, the second data fragment is 192 to 447 of the original data, the third data fragment is 383 to 638 of the original data, and the i-th data fragment is 191*i-190 to 191*i+65 of the original data.

[0015] Step 3: Perform FFT on the data of each slice to obtain the corresponding frequency domain result, and then set the unnecessary frequency components to zero in the frequency domain according to the actual filtering requirements;

[0016] Step 4: multiply the first half of the frequency domain result obtained in step 3 by the complex number i; multiply the second half of the frequency domain result obtained in step 3 by the complex number -i;

[0017] Step 5: Perform IFFT on the result obtained in step 3 to obtain the frequency-domain filtered heterodyne interference signal SIN(D); perform IFFT on the result obtained in step 4 to obtain the orthogonal component COS(D) of the heterodyne interference signal;

[0018] Step 6: The digital processor synchronously generates reference signals SIN(R) and COS(R) of SIN(D) and COS(D), wherein the reference signals SIN(R) and COS(R) are both of data length 2. N The sine and cosine signals have the same frequency as the center frequency of the heterodyne interference signal;

[0019] Step 7: Process SIN(D), COS(D), SIN(R), COS(R) obtained in step 6 as follows:

[0020]

[0021] The processed signals SIN(DR) and COS(DR) are subjected to arc tangent calculation and phase unwrapping:

[0022]

[0023] Among them, kπ is an additional phase, and the value of k is an integer;

[0024] Step 8: Reassemble the data segments of the demodulated phase obtained in step 7, and the length of each data segment is 2 N , the demodulation result is “cut off the head and tail”, that is, only the p5-p6 part of each data segment is taken, N=8, that is, the data of 1-32 and 225-256 are discarded, and only 33-224, a total of 192 data, are taken, and the corresponding data segments are spliced ​​and calculated;

[0025] Step 9: Perform appropriate CIC decimation filtering on the complete demodulated phase to reduce the data rate, and then filter out unnecessary high-frequency components through a digital low-pass filter.

[0026] Furthermore, the specific steps of the splicing calculation in step eight are as follows:

[0027] Step 1: Demodulate data 1 to 192 of group 1, and subtract the value of the 192th data from all data, so that the value of the last data of group 1 is 0;

[0028] Step 2: For the demodulated data 1 to 192 of the second group, the value of the first data is subtracted from all the data, so that the value of the second data is 0; the third group, the fourth group, the fifth group, ... are all processed in the same way as the second group;

[0029] Step 3: Data splicing: splice the complete data of the first group, the 2-192 data of the second group, the 2-192 data of the third group, the 2-192 data of the fourth group, the 2-192 data of the fifth group, ..., in order to obtain the final complete demodulation phase.

[0030] The present invention has the following advantages: the present invention avoids the endpoint effect of the FFT algorithm / Hilbert transform by slicing digital signal data, overlapping calculations and reorganizing demodulation results, and the present invention can demodulate multiple segments of data simultaneously, is particularly suitable for parallel signal processing on a digital processor, and can greatly increase the demodulation efficiency.

[0031] The present invention fully reuses the results of FFT, not only realizing frequency domain filtering, but also efficiently obtaining an orthogonal signal pair of heterodyne interference signals through complex multiplication and IFFT.

[0032] The present invention realizes nearly ideal digital filtering very efficiently by means of zeroing the spectrum. DETAILED DESCRIPTION

[0033] The present invention provides a phase demodulation method for signal slicing and overlapping calculation based on FFT and IFFT frequency domain filtering, and the method comprises the following steps:

[0034] Step 1: Digitally sample the analog heterodyne interference signal. The sampling rate must meet the following requirements:

[0035] f s >2·BW

[0036] Where BW is the signal bandwidth and:

[0037] BW=2(2v p / λ+f v )

[0038] Step 2: Overlap and slice the sampled digital signal. The original data is a continuous one-dimensional digital signal. Slice the one-dimensional data into segments suitable for FFT. The data length must be 2 N , where N is a positive integer), and set the relevant parameters as follows: p1 = 2 N -(2 N / m+1);p2=p1–1;p3=p1;p4=2 N -p2 -1; p5 = (2 N -p1+1) / 2; p6=2N +1-p5; where m is generally 2, 4, 8, or 16. When m=4 and N=8, the first data fragment is 1 to 256 of the original data, the second data fragment is 192 to 447 of the original data, the third data fragment is 383 to 638 of the original data, and the i-th data fragment is 191*i-190 to 191*i+65 of the original data.

[0039] Step 3: Perform FFT on the data of each slice to obtain the corresponding frequency domain result, and then set the unnecessary frequency components to zero in the frequency domain according to the actual filtering requirements;

[0040] Step 4: multiply the first half of the frequency domain result obtained in step 3 by the complex number i; multiply the second half of the frequency domain result obtained in step 3 by the complex number -i;

[0041] Step 5: Perform IFFT on the result obtained in step 3 to obtain the frequency-domain filtered heterodyne interference signal SIN(D); perform IFFT on the result obtained in step 4 to obtain the orthogonal component COS(D) of the heterodyne interference signal;

[0042] Step 6: The digital processor synchronously generates reference signals SIN(R) and COS(R) of SIN(D) and COS(D), wherein the reference signals SIN(R) and COS(R) are both of data length 2. N The sine and cosine signals have the same frequency as the center frequency of the heterodyne interference signal;

[0043] Step 7: Process SIN(D), COS(D), SIN(R), COS(R) obtained in step 6 as follows:

[0044]

[0045] The processed signals SIN(DR) and COS(DR) are subjected to arc tangent calculation and phase unwrapping:

[0046]

[0047] Among them, kπ is an additional phase, and the value of k is an integer;

[0048] Step 8: Reassemble the data segments of the demodulated phase obtained in step 7, and the length of each data segment is 2 N , the demodulation result is “cut off the head and tail”, that is, only the p5-p6 part of each data segment is taken, N=8, that is, the data of 1-32 and 225-256 are discarded, and only 33-224, a total of 192 data, are taken, and the corresponding data segments are spliced ​​and calculated;

[0049] Step 9: Perform appropriate CIC decimation filtering on the complete demodulated phase to reduce the data rate, and then filter out unnecessary high-frequency components through a digital low-pass filter.

[0050] In this embodiment, the specific steps of the splicing calculation in step eight are as follows:

[0051] Step 1: Demodulate data 1 to 192 of group 1, and subtract the value of the 192th data from all data, so that the value of the last data of group 1 is 0;

[0052] Step 2: For the demodulated data 1 to 192 of the second group, the value of the first data is subtracted from all the data, so that the value of the second data is 0; the third group, the fourth group, the fifth group, ... are all processed in the same way as the second group;

[0053] Step 3: Data splicing: splice the complete data of the first group, the 2-192 data of the second group, the 2-192 data of the third group, the 2-192 data of the fourth group, the 2-192 data of the fifth group, ..., in order to obtain the final complete demodulation phase.

[0054] Although the specific implementation of the present invention has been described in detail, it should not be understood as limiting the scope of protection of this patent. Within the scope described in the claims, various modifications and variations that can be made by those skilled in the art without creative work still fall within the scope of protection of this patent.

Claims

1. A phase demodulation method based on signal slicing and overlapping calculation of FFT and IFFT frequency domain filtering, characterized in that: The method comprises the following steps: Step 1: Digitally sample the analog heterodyne interference signal. The sampling rate must meet the following requirements: f s >2·BW Among them, f s is the sampling rate, BW is the signal bandwidth, and: BW=2(2v p / λ+f v ) Among them, v p is the amplitude of the vibration velocity of the object to be measured, λ is the laser wavelength of the laser heterodyne interferometer, and f v is the frequency of the vibration signal; Step 2: Overlap and slice the sampled digital signal. The original data is a continuous one-dimensional digital signal. Slice the one-dimensional data into segments suitable for FFT. The data length must be 2. N , where N is a positive integer, and the relevant parameters are set as follows: p1 = 2 N -(2 N / m+1);p2=p1-1;p3=p1;p4=2 N -p2-1; p5=(2 N -p1+1) / 2; p6=2 N +1-p5; where m is 2, 4, 8, or 16. When m=4 and N=8, the first data fragment is 1 to 256 of the original data, the second data fragment is 192 to 447 of the original data, the third data fragment is 383 to 638 of the original data, and the i-th data fragment is 191*i-190 to 191*i+65 of the original data. Step 3: Perform FFT on the data of each slice to obtain the corresponding frequency domain result, and then set the unnecessary frequency components to zero in the frequency domain according to the actual filtering requirements; Step 4, multiply the first half of the frequency domain result obtained in step 3 by the complex number i; multiply the second half of the frequency domain result obtained in step 3 by the complex number -i; Step 5, performing IFFT on the result obtained in step 3 to obtain the frequency-domain filtered heterodyne interference signal SIN(D); performing IFFT on the result obtained in step 4 to obtain the orthogonal component COS(D) of the heterodyne interference signal; Step 6: The digital processor synchronously generates reference signals SIN(R) and COS(R) of SIN(D) and COS(D), wherein the reference signals SIN(R) and COS(R) are both of data length 2. N The sine and cosine signals have the same frequency as the center frequency of the heterodyne interference signal; Step 7: Process SIN(D), COS(D), SIN(R), COS(R) obtained in step 6 as follows: The processed signals SIN(DR) and COS(DR) are subjected to arc tangent calculation and phase unwrapping: Among them, kπ is an additional phase, and the value of k is an integer; Step 8: Reassemble the data segments of the demodulated phase obtained in step 7. The length of each data segment is 2 N , "cut off the head and tail" of the demodulation result, that is, only take the p5~p6 part of each data segment, N=8, that is, discard the data of 1~32 and 225~256, and only take 33~224, a total of 192 data, and splice the corresponding data segments for calculation; Step 9: Perform appropriate CIC decimation filtering on the complete demodulated phase to reduce the data rate, and then filter out unnecessary high-frequency components through a digital low-pass filter.

2. A phase demodulation method based on signal slicing and overlapping calculation of FFT and IFFT frequency domain filtering as claimed in claim 1, characterized in that: The specific steps of the splicing calculation in step 8 are as follows: Step 81, the demodulated data 1 to 192 of the first group, all data are subtracted from the value of the 192th data, so that the value of the last data of the first group of data is 0; Step 82: For demodulated data 1 to 192 of the second group, the value of the first data is subtracted from all the data, so that the value of the second data is 0; the third group, the fourth group, the fifth group, ... are all processed in the same way as the second group; Step 83, data splicing, splicing the complete data of the first group, the 2-192 data of the second group, the 2-192 data of the third group, the 2-192 data of the fourth group, the 2-192 data of the fifth group, ..., in order to obtain the final complete demodulation phase.

Images