A large-range fast time-stretch sweep-frequency interferometric ranging method based on adaptive frequency shifting

Abstract

The application relates to an absolute distance measurement technology field, and discloses a large-range fast time stretching sweep frequency interferometric ranging method based on adaptive frequency shifting. In order to solve the defects of the prior art, such as the difficulty in balancing the ranging resolution and the range, the influence of the reference frequency drift, the significant decline of the signal-to-noise ratio with the increase of the distance, and the low calculation efficiency of the ranging algorithm, the technical scheme comprises the following steps: acquiring a measurement interference signal and an auxiliary interference signal; performing equal optical frequency interval resampling on the two interference signals; performing mixing and filtering processing on the resampled signals; performing frequency spectrum analysis to obtain a coarse time delay estimation; performing frequency shift correction on the frequency spectrum according to the coarse time delay estimation result; performing cross-correlation analysis on the corrected frequency spectrum to extract a peak position; and combining the coarse time delay estimation result and the cross-correlation analysis result to output a target distance value. The application is suitable for fast, large-range and high-precision absolute distance measurement of non-cooperative targets.

Description

Technical Field

[0001] It involves the fields of absolute distance measurement of non-cooperative targets, fiber optic sensing, OFDR, swept frequency interferometry, OCT, and lidar three-dimensional measurement technology. Background Technology

[0002] Frequency-sweeping interferometric ranging (FSI), an interferometric absolute distance measurement technique, achieves high detection sensitivity through energy accumulation in the frequency domain. It is suitable for non-cooperative measurement scenarios such as low reflectivity, large-angle incidence, and vibration, and can achieve sub-micron measurement accuracy. Furthermore, its absence of distance ambiguity makes it applicable to medium- and long-distance measurements over tens of meters. Currently, 3D imaging LiDAR based on this technology is widely used in the precision assembly of large components, robot calibration and reengineering, and reverse engineering.

[0003] The basic process of frequency sweep interferometric ranging technology is as follows: Figure 1 As shown, the time delay difference is... The two frequency-modulated continuous light beams interfere, and the beat frequency signal received by the photodetector is: (1) in, Indicates the amplitude of the beat signal. This represents the angular frequency of light. It is typically determined using an FFT (Flexible Transformation Principle). Perform quickly Domain spectral analysis, where the computational resolution (range bin) of the distance spectrum is... : (2) It is only related to the sweep bandwidth. In the field of precision measurement with FMCW lidar, due to limitations in laser and data processing efficiency, the sweep bandwidth of the light source is generally <10nm. For precision requirements ranging from tens to even a few micrometers, direct FFT is clearly insufficient. The Zoom FFT using complex modulation-downsampling and the Chirp-Z Transform (CZT) algorithm using spiral sampling are classic methods for obtaining higher frequency resolution within specific narrow spectral ranges. Both require prior information from a pre-executed FFT to plan the frequency range to be subdivided before performing the frequency subdivision operation. Therefore, high time complexity is their main problem. Furthermore, the number of sampling points is limited by the FMCW measurement principle. ( The time complexity of the algorithm is directly proportional to the bandwidth and measurement range. To achieve measurement accuracy at the micrometer or even sub-micrometer level and a ranging range of tens of meters, N typically reaches 100K to 1000K, further exacerbating the algorithm's time complexity. Wang proposed the SCZT algorithm, which performs block-parallel processing on the CZT algorithm to improve processing speed. Based on this, Ma et al. proposed a fast, high-resolution frequency estimation algorithm combining FFT and SCZT. The computation speed at sampling points is 75% faster than the full-size CZT based on the CPU. Algorithm acceleration based on FPGAs and GPUs has also been extensively studied, but the acceleration effect is not significant when the number of sampling points is large. The fundamental reason is that combined algorithms such as FFT+CZT have an unavoidable fundamental flaw: the initial sampling angle phase factor... It needs to be dynamically updated based on the FFT results, and FFT and CZT can only be calculated sequentially; increasing the number of points leads to chirped phase factors. Pre-storage consumes more memory and read latency increases exponentially. Algorithms such as Phase Difference Method, Power Spectral Estimation Method, and Amplitude Ratio Method interpolate based on FFT to improve frequency estimation accuracy, but their accuracy is poor under low signal-to-noise ratio conditions and they can only be used for single targets.

[0004] Yongwoo Park et al. proposed a cross-chirped interferometry system and method for ranging. However, this method has a serious problem: the signal-to-noise ratio roll-off severely limits the ranging range. For example... Figure 2 As shown, the spectral width of the interference signal is limited by the dispersive medium and the maximum bandwidth B of the swept frequency light source. As the target distance increases, the overlap between the spectrum of the reference signal and the spectrum of the measurement signal decreases, which leads to a gradual decrease in the amplitude of the correlation result. This causes the signal-to-noise ratio of the correlation spectrum to roll off as the measured distance increases, which seriously affects the ranging range.

[0005] In summary, existing technologies suffer from drawbacks such as difficulty in balancing ranging resolution and range, significant impact of reference frequency drift, a substantial decrease in signal-to-noise ratio with increasing distance, and low computational efficiency of ranging algorithms. Summary of the Invention

[0006] To address the shortcomings of existing technologies, such as the difficulty in balancing ranging resolution and range, significant impact of reference frequency drift, substantial decrease in signal-to-noise ratio with increasing distance, and low computational efficiency of ranging algorithms, the technical solution provided by this invention is as follows: A large-range, fast time-stretching frequency sweep interferometric ranging method based on adaptive frequency shifting includes: The steps for acquiring the measurement interference signal and the auxiliary interference signal; The steps for resampling the two interference signals at equal optical frequency intervals; The steps of mixing and filtering the resampled signal; The steps to perform spectrum analysis and obtain a coarse time delay estimate; The steps for frequency shifting correction of the spectrum based on coarse time delay estimation results; The steps include performing cross-correlation analysis on the corrected spectrum to extract peak positions; The steps to output the target distance value are based on the coarse time delay estimation results and the cross-correlation analysis results.

[0007] Furthermore, a preferred embodiment is provided in which the measured interference signal and the auxiliary interference signal originate from a common-source double interference structure.

[0008] Furthermore, a preferred implementation is provided in which the resampling operation is performed based on an equal-frequency interval clock provided by the resampling interference signal.

[0009] Furthermore, a preferred embodiment is provided in which the mixing process includes performing digital mixing on the measured interference signal and the auxiliary interference signal, and extracting the envelope signal through a low-pass filter.

[0010] Furthermore, a preferred implementation is provided in which the coarse time delay estimation result is obtained by detecting narrowband peaks in the spectrum of the measured signal.

[0011] Based on the same inventive concept, the present invention also provides a large-range fast time-stretching frequency sweep interferometric ranging device based on adaptive frequency shifting, comprising: A module for acquiring the measurement interference signal and the auxiliary interference signal; A module for resampling two interference signals at equal optical frequency intervals; A module that performs mixing and filtering on the resampled signal; A module that performs spectrum analysis to obtain coarse time delay estimates; A module that performs frequency shift correction on the spectrum based on coarse time delay estimation results; A module for performing cross-correlation analysis on the corrected spectrum and extracting peak positions; This module outputs the target distance value by combining the coarse time delay estimation results with the cross-correlation analysis results.

[0012] Based on the same inventive concept, the present invention also provides a ranging system for implementing the method, comprising: A swept-frequency laser is used to emit linearly frequency-modulated optical signals; An optical splitter structure is used to divide an optical signal into three paths; The time stretching module is used to perform dispersion stretching on two reference beams; The measuring interferometer, the auxiliary interferometer, and the resampling auxiliary interferometer each receive the split optical signal. The measuring interferometer and the auxiliary interferometer form a beat frequency interference signal through the time stretching module, and the resampling auxiliary interferometer provides an equal optical frequency interval interference signal. The detector module is used to receive the interference signals output by each interferometer and convert them into electrical signals.

[0013] Based on the same inventive concept, the present invention also provides a computer storage medium for storing a computer program, wherein when the computer program is read by a computer, the computer executes the method described thereon.

[0014] Based on the same inventive concept, the present invention also provides a computer, including a processor and a storage medium, wherein when the processor reads a computer program stored in the storage medium, the computer executes the method described thereon.

[0015] Based on the same inventive concept, the present invention also provides a computer program product, which, when executed, implements the method described.

[0016] Compared with the prior art, the advantages of the technical solution provided by the present invention are as follows: By introducing a dual-optical time-stretching interference structure of the same origin, the reference light and the measurement light share the same swept-frequency laser source and undergo consistent optical path changes, thereby constructing a structurally symmetrical and signal-matched interference channel. Compared with traditional FSI systems where spectral analysis heavily relies on serial processing algorithms such as FFT+CZT and has low overall computational efficiency, this structure provides the foundation for parallelization and structural simplification of subsequent adaptive frequency shifting and cross-correlation processing, significantly improving the system's computational efficiency and processing speed, and solving the problem of slow response speed of traditional FSI algorithms in large-scale ranging scenarios.

[0017] To address the limitation of traditional cross-correlation ranging methods in long-distance scenarios, this scheme introduces a non-equilibrium time-stretched interferometric structure. Dispersive elements are added to the reference arm, transforming the interference signal from a single-frequency beat signal into a chirped signal, thus converting the frequency estimation problem into a cross-correlation time delay estimation problem. By mapping the time delay to a narrow pulse position in the frequency domain, high-resolution localization of the signal's main lobe is achieved. Compared to the cross-chirped interferometric method proposed by Yongwoo Park et al., which suffers from narrowed spectral overlap regions and difficulty in effective demodulation at long distances, this scheme significantly expands the effective ranging range of the cross-correlation algorithm by zooming the cross-correlation spectrum and combining it with a resampling mechanism. This overcomes the trade-off between computational window and frequency resolution that limits the ranging capability of traditional methods.

[0018] This scheme, based on the measurement interferometer and auxiliary interferometer, proposes to obtain the k-domain resampling clock through a resampling interferometer to correct the nonlinear characteristics of the swept-frequency laser. This design solves the distance calculation error problem caused by the nonlinearity of the frequency-time mapping in traditional systems, achieves resampling with equal optical frequency intervals, and improves the overall measurement accuracy of the system. Compared with the FSI system without resampling correction, the spectral distribution is more uniform, and the distance spectrum resolution obtained after FFT operation is higher.

[0019] The solution further incorporates an Adaptive Frequency-Shifted Cross-Correlation Demodulation (AFS-CC) algorithm to expand the effective ranging range of the cross-correlation algorithm. Its core lies in using mixing and frequency shifting operations to ensure the measured signal and reference signal coincide in the frequency domain, thereby expanding the effective range of the frequency domain cross-correlation and solving the problem of the reduced ranging range caused by zoom limitations in traditional cross-correlation methods. Compared to existing FFT+CZT combined algorithms, AFS-CC can achieve unambiguous high-precision ranging at the 50-meter level without sacrificing resolution.

[0020] In terms of computational efficiency, this algorithm avoids the complex complex factor calculations and phase updates of CZT by utilizing a parallel structure and frequency domain sparsity, significantly reducing the number of FFTs and memory access latency. Compared to the FFT+SCZT algorithm proposed by Ma et al., this scheme can reduce the FFT computation by at least one-third and demonstrates a 447% computational speedup in simulations, greatly improving processing efficiency under large point count conditions and making it more suitable for real-world long-distance real-time ranging scenarios.

[0021] This scheme truncates the chirped interference signal in the frequency domain and replaces the full-point FFT with a small-point DFT, greatly reducing the number of complex multiplications in the data processing. Compared with existing Zoom-FFT or CZT methods that require full-range FFT processing, its sparse frequency domain analysis structure significantly reduces the computational burden while maintaining equivalent resolution, making the algorithm more lightweight, engineering-friendly, and adaptable.

[0022] It is suitable for high-precision, large-range non-cooperative target absolute distance measurement, especially in scenarios such as lidar 3D modeling, precision assembly and industrial measurement. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the basic FSI ranging process; Figure 2 This is a schematic diagram illustrating the causes of signal-to-noise ratio roll-off. Figure 3 This is a schematic diagram illustrating the principle of a large-range ranging method based on dual optical time-stretched frequency sweeping interferometry of the same origin. Figure 4 This is a schematic diagram of the optical path structure of the ranging system; Figure 5 This is a schematic diagram of the time stretching module; Figure 6 Flowchart of the full-range distance demodulation algorithm for adaptive frequency shift cross-correlation; Figure 7 This is a schematic diagram of the simulation test results; Figure 8 This diagram illustrates the computation time consumption of FFT, AFS-CC, and FFT+CZT. Detailed Implementation

[0024] To make the advantages and benefits of the technical solution provided by the present invention clearer, the technical solution provided by the present invention will now be described in further detail with reference to the accompanying drawings, specifically: Implementation Method 1: This implementation method provides a large-range, fast time-stretching frequency sweep interferometric ranging method based on adaptive frequency shifting, including: The steps for acquiring the measurement interference signal and the auxiliary interference signal; The steps for resampling the two interference signals at equal optical frequency intervals; The steps of mixing and filtering the resampled signal; The steps to perform spectrum analysis and obtain a coarse time delay estimate; The steps for frequency shifting correction of the spectrum based on coarse time delay estimation results; The steps include performing cross-correlation analysis on the corrected spectrum to extract peak positions; The steps to output the target distance value are based on the coarse time delay estimation results and the cross-correlation analysis results.

[0025] The measured interference signal and the auxiliary interference signal are derived from the same source double interference structure.

[0026] The resampling operation is based on an equal-frequency interval clock provided by the resampling interference signal.

[0027] The mixing process includes digitally mixing the measured interference signal with the auxiliary interference signal and extracting the envelope signal through a low-pass filter.

[0028] The coarse time delay estimation result is obtained by detecting narrowband peaks in the spectrum of the measured signal.

[0029] A ranging system is also provided for implementing the method, comprising: A swept-frequency laser is used to emit linearly frequency-modulated optical signals; An optical splitter structure is used to divide an optical signal into three paths; The time stretching module is used to perform dispersion stretching on two reference beams; The measuring interferometer, the auxiliary interferometer, and the resampling auxiliary interferometer each receive the split optical signal. The measuring interferometer and the auxiliary interferometer form a beat frequency interference signal through the time stretching module, and the resampling auxiliary interferometer provides an equal optical frequency interval interference signal. The detector module is used to receive the interference signals output by each interferometer and convert them into electrical signals.

[0030] Implementation Method Two: This implementation method is a further explanation of the technical solution provided in Implementation Method One, specifically: To achieve an absolute distance measurement method suitable for high resolution, large range, and strong anti-interference capability, the following technical solution is proposed, the execution process of which includes the following steps: Step 1: Construction and Optical Path Configuration of the Same-Source Double Interference System First, a frequency-sweeping interferometric ranging system based on a co-originating dual-interference structure is constructed. This system comprises three parts: a tunable laser source, a measuring interferometer, an auxiliary interferometer, and a resampling interferometer. These parts are interconnected by optical components to form a fixed topology. The three interferometers share the same laser and a portion of the optical path to ensure initial phase and frequency synchronization between the reference signal and the measured signal.

[0031] In the measurement interferometer, the laser is output to the measuring arm and the reference arm via a beam splitter. The measuring arm is focused onto the target by a lens and then returns. Dispersive fiber is introduced into the reference arm to form a time-stretched path. The combined light from both is received by a balanced detector and outputs a chirped beat frequency signal.

[0032] The auxiliary interferometer is used to capture the initial frequency fluctuation information of the same source light source. Its structure is basically the same as that of the measurement interferometer, but it does not include the target reflection path. Instead, it constructs a fixed time delay difference structure through fiber delay lines.

[0033] The resampling interferometer is used to provide an equal-frequency sampling clock signal. Its measuring arm and reference arm have a stable and known time delay difference, forming an interference signal source with a fixed frequency interval.

[0034] The output of this step is three interference signals, which are used for distance information acquisition, drift compensation, and nonlinear correction, respectively.

[0035] Step Two: Acquisition and Preliminary Conditioning of Interference Signals Next, a high-speed data acquisition card is used to synchronously acquire the measurement interference signal, auxiliary interference signal, and resampled interference signal. To improve system accuracy, the sampling clock of the acquisition card is not determined by the internal clock, but by providing an external trigger clock through the output of the resampled interference signal after passing through a comparator and conditioning circuit, thus achieving k-space equidistant resampling.

[0036] The resampling operation converts the three interferometric signals into an interferometric data sequence with a uniform optical frequency spacing, correcting the nonlinear distortion during the frequency sweeping process of the swept laser. The output of this step is the resampled interferometric signal sequence, which has high frequency consistency.

[0037] Step 3: Complex modeling and mixing of the interference signal After sampling and resampling, the measured interference signal and the auxiliary interference signal are modeled as chirped signals in complex modulation form, containing unbalanced delay information where the angular frequency varies with time. Since the initial frequency drift affects the synchronization of the two signals, the two signals are digitally mixed to eliminate common-mode initial frequency drift.

[0038] The mixed signal is low-pass filtered to obtain the envelope signal, which then enters the spectrum analysis stage. The output of this step is a stable interference signal with frequency drift removed and its spectrum representation.

[0039] Step 4: Adaptive Frequency Shift Cross-Correlation Calculation Based on FFT The two processed interference signals are subjected to Fast Fourier Transform (FFT) to obtain their spectra. A high-amplitude narrow pulse exists in the spectrum of the measured signal; peak detection of this pulse can yield an initial coarse time delay difference.

[0040] The frequency domain shift operation is performed on the spectrum of the measured signal based on the estimated time delay difference. This involves truncating a certain number of points to the left of the spectrum and filling the end of the spectrum with the same number of zero values ​​while keeping the total number of points unchanged, thereby minimizing the residual time delay difference between the two signals.

[0041] The output of this step is the frequency-shift corrected spectrum of the measurement signal, which satisfies the prerequisite for cross-correlation precision measurement.

[0042] Step 5: Refined Cross-Correlation Analysis and Distance Spectrum Construction The cross-correlation spectrum is obtained by multiplying the spectrum of the corrected measurement signal with the spectrum of the auxiliary signal by complex conjugate. The spectrum is then clipped to retain the non-zero part, and a small-point discrete Fourier transform (DFT) is performed to output a high-resolution cross-correlation spectrum.

[0043] In the cross-correlation spectrum, the peak position corresponds to the optical path delay. The distance to the target is calculated using this peak position to obtain a high-precision ranging result. The output of this step is a narrow pulse result processed by zoom cross-correlation, which can be used for sub-micron level time delay estimation.

[0044] Step Six: Fusion and Decoding of Coarse and Fine Measurement Results By fusing the distance corresponding to the coarse time delay measured in step four with the high-precision distance calculated from the cross-correlation peak position obtained in step five, and automatically matching the closest integer cycle distance value through an adaptive frequency shift demodulation algorithm, a wide-range, unambiguous, and high-precision ranging measurement can be achieved.

[0045] The output of this step is the final high-precision distance value, which is suitable for measuring the absolute position of targets within a ranging range of 0 to 50 meters.

[0046] Step 7: Efficiency Optimization and System Execution Environment Deployment To further improve algorithm efficiency, the sparsity of chirped signals in the frequency domain is utilized to implement frequency domain truncation and replace large-point FFT with small-point DFT; and a GPU parallel acceleration architecture is adopted to process spectrum calculation and mixing operation in parallel, further reducing the overall computational latency.

[0047] This step outputs the system deployment structure and real-time execution environment parameter configuration, demonstrating engineering feasibility and high-performance implementation capabilities.

[0048] Regarding the optical path structure: Figure 4 The overall optical path structure of the ranging system proposed in this invention is shown. The entire system mainly consists of the following parts: 1. TLS (Tunable Laser Source) – Frequency-swept laser Function: Outputs a linearly frequency-modulated, narrow-linewidth laser signal as the unified light source for the system; Requirements: Wide sweep bandwidth and good frequency stability; The output light is simultaneously fed into three interferometers as an input signal.

[0049] 2. Optical Coupler System Function: Splits the optical signal output by TLS into three paths: The signal is fed into the measuring interferometer. The feed is sent directly into the auxiliary interferometer; One path is fed into the resampling-assisted interferometer; Requirements: Consistent splitting ratio, reasonable distribution of signal strength, and all three beams must originate from the same source.

[0050] 3. Time-stretching Module Function: Used to introduce group delay dispersion in the reference optical path, stretching the original optical signal in the time domain to achieve modulation of the chirped signal.

[0051] Location: Located in the reference arm of the measuring interferometer and the auxiliary interferometer, sharing this module.

[0052] Implementation: See details for the specific structure. Figure 5 This can be achieved in two ways.

[0053] 4. Measurement Interferometer Structure: It consists of a measuring arm and a reference arm, which are ultimately combined at the photodetector to form an interference signal; Measuring arm path: Optical signal → OC1 → CIR1 (optical circulator) → Lens → Target (object under test) reflection; Echo → Lens → CIR1 → OC4 → BPD1; Reference arm path: Optical signal → OC1 → OC2 → CIR2 → CFBG (dispersion grating) → CIR2 → OC5 → OC4 → BPD1; Features: The non-equilibrium interference structure is formed by stretching only the reference arm; The output beat frequency signal (a chirped signal) is input to BPD1 for use in main distance demodulation.

[0054] 5. Auxiliary Interferometer Function: To generate a cross-correlation beat frequency signal with the measuring interferometer, used to eliminate initial frequency drift error; Measuring arm path: TLS → OC1 → OC2 → OC3 → DL1 (Delayed Fiber) → OC6 → BPD2; Reference arm path: TLS → OC1 → OC2 → CIR2 → CFBG → CIR2 → OC5 → OC6 → BPD2; Features: The two interfering signals are of the same origin and are synchronous, and the initial frequency fluctuation has a common-mode effect; Cross-correlation can cancel common-mode error and enhance system stability; Output beat frequency signal input BPD2.

[0055] 6. Resampling Auxiliary Interferometer Function: Generates equal-frequency-interval interference signals and outputs an external clock signal for k-domain resampling of beat frequency signals; Measuring arm path: TLS → OC1 → OC2 → OC3 → OC7 → DL2 → OC8 → BPD3; Reference arm path: TLS → OC1 → OC2 → OC3 → OC7 → OC8 → BPD3; Features: Interference fringes are formed using delay fiber DL2; The output is used for an optical frequency clock signal; It provides a precise resampling reference, solving the problem of nonlinear frequency modulation of the light source; The output signal is input to BPD3 and sent to a comparator or shaping circuit as an external clock for the data acquisition card.

[0056] 7. BPD1 / BPD2 / BPD3 (Balanced Photodetector) Function: Used to detect the output optical signal of a three-way interferometer and convert it into an electrical signal; Requirements: High response rate and low noise, for accurate capture of interferometric beat frequencies.

[0057] Figure 5 Detailed Explanation of Two Implementation Methods for the Time Stretching Module The time stretching module is one of the core modules of this system. Its purpose is to introduce group delay dispersion on the reference optical path, generate a chirped signal, and improve the distance demodulation resolution.

[0058] Implementation Method 1: Based on CFBG (Chired Fiber Bragg Grating) structure: Input light → CIR (optical circulator) → CFBG → CIR output; effect: CFBG provides fixed group delay dispersion (GDD), which allows for precise control of the frequency-time mapping. The stretched reference light exhibits a linear frequency change in the frequency domain.

[0059] With a compact structure and high stability, it is suitable for conventional fixed dispersion requirements, has high linearity during stretching, and can be customized to 0 for higher-order dispersion.

[0060] Implementation Method 2: Based on DCF (Dispersion Compensating Fiber) structure: Input light → CIR → DCF → out effect: DCF dispersion-compensated fiber serves the same purpose as CFBG, but with advantages such as lower cost, easier adjustment of dispersion, and the ability to achieve greater bandwidth and group delay dispersion.

[0061] Figure 4 The core innovation of this system, the three-way interference structure and its optical connection method, are demonstrated. Figure 5 The implementation method of the time stretching module is specifically given. The entire structure effectively solves key technical problems such as initial frequency drift of the light source, nonlinear frequency modulation and zero-frequency noise interference through homogeneous design, double interference structure, dispersion stretching and resampling mechanism, and realizes a high-precision, robust and wide-range high-performance absolute ranging system.

[0062] Implementation Method 3: Combination Figure 1-8 This embodiment describes the technical solution provided above in further detail through specific examples. Specifically: A large-range ranging method based on homogeneous dual-optical time-stretched sweeping interferometry, the idea of ​​which is as follows: Figure 3 As shown, a dispersive medium is first introduced onto the reference arm to obtain a local oscillator light (stretched light) with a changed frequency modulation slope. This local oscillator light, together with the unstretched measurement light, forms a non-equilibrium time-stretched interference. The beat signal changes from a single-frequency signal in FSI to a chirped signal, and the frequency estimation problem can be transformed into a correlation delay estimation problem for the chirped signal. The group delay dispersion of the dispersive medium... Mapping the target time delay to the optical frequency domain, and... Figure 3 The peak positions of the cross-correlation spectrum correspond to each other, because Extremely small The spectral line spacing FSR corresponding to the domain Much smaller than the spectral line spacing in FFT Therefore, this method can more precisely characterize the cross-correlation spectrum and improve the accuracy of time delay estimation. This process can be figuratively described as starting from... Domain FFT spectrum to Domain cross-correlation spectrum zoom magnification, zoom magnification is The system employs a dual-optical time-stretched sweep frequency interference structure of the same origin to ensure synchronization between the measurement signal and the reference signal, and suppresses the initial frequency drift of the light source through cross-correlation.

[0063] Based on this ranging method, the ranging system is as follows: Figure 4 As shown, the system consists of three parts: a measurement interferometer, an auxiliary interferometer, and a resampling auxiliary interferometer. The measurement interferometer and the auxiliary interferometer share a time stretching module, which can... Figure 5 The two methods shown are used for construction; Specifically, the measurement interferometer consists of the measurement arm TLS-OC1-CIR1-Lens-Target-Lens-CIR1-OC4-BPD1 and the reference arm TLS-OC1-OC2-CIR2-CFBG-CIR2-OC5-OC4-BPD1; the auxiliary interferometer consists of... Measuring arm TLS-OC1-OC2-OC3-DL1-OC6-BPD2 The system consists of a reference arm TLS-OC1-OC2-CIR2-CFBG-CIR2-OC5-OC6-BPD2 and a resampling interferometer measuring arm TLS-OC1-OC2-OC3-OC7-DL2-OC8-BPD3. The measuring interferometer and auxiliary interferometer share a single TLS light source and the same time-stretching module, forming a co-source dual-interference structure to suppress the influence of initial wavelength fluctuations of the light source on the ranging results. Since the sweep frequency and time do not satisfy a strict linear relationship, a method of equal optical frequency interval resampling (k-domain resampling) is generally used to resample the aforementioned non-uniform beat signal. The resampling clock comes from the interference signal of the resampling interferometer, and the time delay difference between the measuring arm and the reference arm satisfies the following condition: The optical frequency resampling process can be represented as: (3) Therefore, the total number of sampling points can be obtained. ,in The relative change in the angular frequency of the light source is given by the free spectral range (FSR) of the resampling interferometer. .

[0064] Beat frequency signal generated by measuring interferometer It can be expressed in complex form as: (4) in, This represents the change in angular frequency, with an initial value of 0 and a final value of [missing value]. , For the sweep bandwidth, This indicates the time delay difference of the measuring interferometer. Assuming the initial angular frequency of the swept beam is constant and the center frequency of the CFBG is constant, let the initial angular frequency of the swept beam be constant. The difference is This value fluctuates during repeated frequency sweeps. This represents the change in time delay introduced by the initial optical frequency fluctuation, which will lead to... The distance drift error severely degrades the repeatability and stability of the ranging system.

[0065] To suppress To address the reference drift caused by fluctuations, an additional non-equilibrium time-stretching auxiliary interferometer of the same origin was designed to mitigate the fluctuations in the initial optical frequency. It remains basically synchronized with the measurement signal. The auxiliary interferometer's beat signal can be expressed as: (5) in, To assist in understanding the time delay difference between the interferometer's measuring arm and the reference arm, pre-calibration can be performed before the measurement begins. The definitions of the remaining symbols are consistent with those in equation (4).

[0066] The beat signal of the resampling interferometer is used as an external clock for the acquisition card through conditioning circuits such as comparators, and the clock signal is used to adjust the two interference signals mentioned above. and To correct the nonlinearity of the swept-frequency laser, k-domain equal-interval resampling is performed. Substituting equation (3) into equations (4) and (5), the resampled signal can be obtained: (6) (7) Cross-correlation of the resampled signals yields a compressed narrow pulse output. : (8) Time delay error caused by initial optical frequency fluctuations in the two interference signals Since they are essentially the same and are common-mode terms, the time delay error in the two interfering signals can be reduced through cross-correlation processing. They cancel each other out, thus suppressing reference drift.

[0067] Finally, searching peak position The distance to be measured can be obtained. : (9) The ranging formula shows that it can suppress the reference drift problem caused by the initial optical frequency fluctuation of the swept frequency light source, and greatly improve the repeatability of the measurement system.

[0068] According to equation (9), the calculation resolution (Range bin) of the cross-correlation distance spectrum can be obtained. : (10) It can be seen that reducing the FSR or GDD of CFBG can effectively improve the computational resolution. To achieve a detection range of tens of meters, the length of the resampling interferometer is generally 100m-200m, hence the FSR is in the sub-MHz range, and the GDD of CFBG is usually in the hundreds. Therefore, this method can easily achieve submicron computational resolution.

[0069] Wherein, FSR represents the free spectral range, CFBG represents the chirped fiber Bragg grating, and GDD represents group delay dispersion.

[0070] from arrive The zoom magnification is: (11) However, zoom magnification also proportionally reduces the maximum measurement distance of this method: (12) Therefore, to achieve medium-to-long-distance measurements at the m level, further expansion is needed on this basis.

[0071] This paper proposes an adaptive frequency-shift cross-correlation full-range range demodulation algorithm (AFS-CC) without adding any components, solving the problem of small ranging range of direct cross-correlation. Furthermore, the algorithm is accelerated based on the frequency domain sparsity of the interferometric signal. The algorithm flow is as follows: Figure 6 As shown; The algorithm employs a parallel structure of FFT + cross-correlation. The algorithm flow is described in detail below: (1) To and The signal is mixed, and then passed through a digital low-pass filter to obtain... .

[0072] (2) At the same time , and Perform a Fourier transform (FFT) to obtain their respective spectra.

[0073] (3) Estimation of the number of frequency shift points in the frequency domain. The spectrum contains a narrow pulse with a high amplitude, such as Figure 7 As shown, the x-coordinate corresponding to its maximum value is and The time delay difference between two signals can be estimated by finding the peak of a narrow pulse. Delay difference estimate The corresponding number of points is .

[0074] (4) Spectrum shifting. The spectrum is truncated to the left. 1 data point, and perform at the end of the spectrum. Zero-padding is used to keep the total number of points N constant. The frequency-shifted signal is denoted as... , and The remaining delay difference is .

[0075] (5) Inverse Fourier Transform. [The following is a partial translation of the original text:] ... The FFT result is conjugate, and in The FFT results are multiplied, and the product sequence is then cut to retain only the spectral sequence of the non-zero region. The cut sequence is then subjected to DFT calculation with L-point input and M-point output to obtain the cross-correlation spectrum.

[0076] (6) Locate the peaks in the cross-correlation spectrum and record the peak positions as... Substituting into equation (9), we can obtain corresponding distance , This indicates the optical path difference within the optical fibers of the two fibers.

[0077] (7) Combining the coarse measurement results from step 3 and the fine measurement results from step 6, the final distance measurement result can be obtained: .

[0078] The ranging range of this method is limited only by the sampling bandwidth, and it has the advantages of large range and no ambiguity. It can easily achieve absolute distance measurement within a range of 50m, which can be expressed as: (13) Simulations were performed on the target object in the range of 0 to 50 meters. Figure 7 Indicates the preliminary measurement result Precision measurement results And measurement error. From Figure 7 It can be seen that the error of the coarse measurement is within the range of c / 2B, which also determines that the range of the fine measurement result is (-c / 2B, c / 2B). After combining the coarse and fine measurement results, the measurement error is less than 0.1μm, and the simulation result is consistent with the theoretical accuracy. (14) In terms of efficiency, compared with the FFT+CZT algorithm in traditional FSI, the adaptive frequency shift cross-correlation algorithm is optimized in the following three aspects: First, it avoids complex phase factors. Dynamic computation and access reduce latency and resource consumption; secondly, the first three FFTs do not require prior knowledge and can be synchronized through parallel processing, greatly improving signal processing speed; finally, since the Fourier transform results of chirped signals are sparsity, a small-point DFT can replace the full-point IFFT, further reducing the computational load. The effective signal length L in the frequency domain satisfies the following equation: (15) The deviation of the rough measurement results satisfies Therefore, the size of the interval to be zoomed in is The signal length M within the interval is equal to the zoom factor. .because Therefore, the number of complex multiplications in the clipped DFT is much smaller than that in the full-point FFT.

[0079] The meaning of "zoom in" is to perform "zoom amplification" processing only on a smaller frequency band determined by the coarse measurement results.

[0080] To evaluate the demodulation speed improvement of the AFS-CC algorithm compared to the FFT+CZT algorithm, the following simulation was conducted. The test environment was as follows: NVIDIA GeForce RTX 3070 Ti Laptop GPU, 12th Gen Intel(R) Core i7-12700H (2.70 GHz), CUDA Version 12.0. The simulation parameters GDD remained unchanged. Different sampling points were obtained by changing the sweep period. The number of points M and L can be obtained by formulas (15) and (11). It should be noted that this simulation only evaluates the algorithm efficiency. Therefore, only the GPU computation time is analyzed. The memory copy time of data transfer from the host to the device (e.g., GPU) and from the device back to the host is not included in the final computation time statistics. The test results are as follows. Figure 7 As shown.

[0081] like Figure 8 When the number of sampling points N is small, the computation time of AFS-CC is close to that of a single FFT. As the number of sampling points increases, it gradually reaches the GPU performance limit, at which point the former's computation time is about three times that of the latter, corresponding to the first three FFTs of AFS-CC. Therefore, with sufficient computing power, the computation speed can be further improved. Compared with FFT+CZT, the computation speedup of AFS-CC increases with the number of sampling points. When the number of sampling points reaches 1,048,576, the computation speedup is increased by 447%, without affecting the chirped phase factor. The timing (pre-stored) indicates that the actual speedup is better than the current test results.

[0082] The advantages of this implementation method are: A large-range distance demodulation algorithm with adaptive frequency shift cross-correlation is proposed, which makes the time delay difference between the measured signal and the reference signal close to the desired value through adaptive frequency shifting. This method solves the problem of signal-to-noise ratio roll-off caused by increasing target distance, greatly improves the signal-to-noise ratio of the cross-correlation signal, increases the effective ranging range of the method, and breaks the contradiction between computational resolution and ranging range.

[0083] In terms of efficiency, compared with the FFT+CZT algorithm in traditional FSI, the adaptive frequency shift cross-correlation algorithm is optimized in the following three aspects: First, it avoids complex phase factors. The algorithm achieves several improvements: firstly, dynamic computation and storage reduce latency and resource consumption in digital systems; secondly, the first three FFTs require no prior knowledge and can be synchronized in parallel, significantly improving signal processing speed; and finally, it utilizes the frequency domain sparsity of chirped interference signals to cut the frequency domain signal, reducing algorithm complexity. In summary, this algorithm improves computation speed by at least 440%.

[0084] The above description of several specific embodiments further details the technical solution provided by the present invention in order to highlight the advantages and benefits of the technical solution provided by the present invention. However, the above-described specific embodiments are not intended to limit the present invention. Any reasonable modifications and improvements to the present invention, combinations of embodiments, and equivalent substitutions based on the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A large-range, fast time-stretching frequency sweep interferometric ranging method based on adaptive frequency shifting, characterized in that, include: The steps for acquiring the measurement interference signal and the auxiliary interference signal; The steps for resampling the two interference signals at equal optical frequency intervals; The steps of mixing and filtering the resampled signal; The steps to perform spectrum analysis and obtain a coarse time delay estimate; The steps for frequency shifting correction of the spectrum based on coarse time delay estimation results; The steps include performing cross-correlation analysis on the corrected spectrum to extract peak positions; The steps for outputting the target distance value by combining the coarse time delay estimation results and the cross-correlation analysis results; Specifically: Beat frequency signal generated by the measuring interferometer Beat frequency signal with auxiliary interferometer The signal is mixed, and then passed through a digital low-pass filter to obtain... ; At the same time , and Perform Fourier transform to obtain their respective spectra; Perform frequency domain frequency shift point estimation. The spectrum contains a narrow pulse with a high amplitude, and the horizontal axis corresponding to its maximum value is... and The time delay difference between two signals can be estimated by finding the peak of a narrow pulse. Delay difference estimate The corresponding number of points is ; This indicates the time delay difference of the measuring interferometer. To assist the interferometer in measuring the time delay difference between the measurement arm and the reference arm, FSR represents the free spectral range; Perform spectrum shifting, truncating the spectrum to the left. 1 data point, and perform at the end of the spectrum. Zero-padding is used to keep the total number of points N constant. The frequency-shifted signal is denoted as... , and The remaining delay difference is ; Perform inverse Fourier transform, Take the conjugate of the FFT result and with The FFT results are multiplied together, and the product sequence is then cut to retain only the spectral sequence of the non-zero region. The cut sequence is then subjected to DFT calculation with L-point input and M-point output to obtain the cross-correlation spectrum. Find the peaks in the cross-correlation spectrum and denote the peak positions as... Substitution , The distance to be measured is... for peak position, This is a compressed, narrow pulse output. This represents the group delay dispersion of the dispersive medium in the time-stretched module, where c represents the speed of light in a vacuum. Indicates the refractive index of air. To assist in measuring the optical path difference between the measuring arm and the reference arm of the interferometer within the optical fiber, achievable corresponding distance , The computational resolution of the cross-correlation distance spectrum; for The corresponding distance; Combining the preliminary measurement results and the detailed measurement results The final ranging result is obtained: , This represents the number of frequency shift points determined based on the coarse time delay estimate. This represents the coarse distance measured based on the frequency shift of the spectrum obtained from the coarse time delay estimation. The measured interference signal and the auxiliary interference signal are derived from the same source double interference structure.

2. The large-range fast time-stretching frequency sweep interferometric ranging method based on adaptive frequency shifting according to claim 1, characterized in that, The resampling operation is based on an equal-frequency interval clock provided by the resampling interference signal.

3. The large-range fast time-stretching frequency sweep interferometric ranging method based on adaptive frequency shifting according to claim 1, characterized in that, The mixing process includes digitally mixing the measured interference signal with the auxiliary interference signal and extracting the envelope signal through a low-pass filter.

4. The large-range fast time-stretching frequency sweep interferometric ranging method based on adaptive frequency shifting according to claim 1, characterized in that, The coarse time delay estimation result is obtained by detecting narrowband peaks in the spectrum of the measured signal.

5. A large-range, fast time-stretching frequency sweep interferometric ranging device based on adaptive frequency shifting, characterized in that, include: A module for acquiring the measurement interference signal and the auxiliary interference signal; A module for resampling two interference signals at equal optical frequency intervals; A module that performs mixing and filtering on the resampled signal; A module that performs spectrum analysis to obtain coarse time delay estimates; A module that performs frequency shift correction on the spectrum based on coarse time delay estimation results; A module for performing cross-correlation analysis on the corrected spectrum and extracting peak positions; A module that combines coarse time delay estimation results with cross-correlation analysis results to output the target distance value; Specifically: Beat frequency signal generated by the measuring interferometer Beat frequency signal with auxiliary interferometer The signal is mixed, and then passed through a digital low-pass filter to obtain... ; At the same time , and Perform Fourier transform to obtain their respective spectra; Perform frequency domain frequency shift point estimation. The spectrum contains a narrow pulse with a high amplitude, and the horizontal axis corresponding to its maximum value is... and The time delay difference between two signals can be estimated by finding the peak of a narrow pulse. Delay difference estimate The corresponding number of points is ; This indicates the time delay difference of the measuring interferometer. To assist the interferometer in measuring the time delay difference between the measurement arm and the reference arm, FSR represents the free spectral range; Perform spectrum shifting, truncating the spectrum to the left. 1 data point, and perform at the end of the spectrum. Zero-padding is used to keep the total number of points N constant. The frequency-shifted signal is denoted as... , and The remaining delay difference is ; Perform inverse Fourier transform, Take the conjugate of the FFT result and with The FFT results are multiplied together, and the product sequence is then cut to retain only the spectral sequence of the non-zero region. The cut sequence is then subjected to DFT calculation with L-point input and M-point output to obtain the cross-correlation spectrum. Find the peaks in the cross-correlation spectrum and denote the peak positions as... Substitution , The distance to be measured is... for peak position, This is a compressed, narrow pulse output. This represents the group delay dispersion of the dispersive medium in the time-stretched module, where c represents the speed of light in a vacuum. Indicates the refractive index of air. This indicates the optical path difference between the measuring arm and the reference arm of the auxiliary interferometer inside the optical fiber. have to corresponding distance , The computational resolution of the cross-correlation distance spectrum; Combining the preliminary measurement results and the detailed measurement results The final ranging result is obtained: , This represents the number of frequency shift points determined based on the coarse time delay estimate. This represents the coarse distance measured based on the frequency shift of the spectrum obtained from the coarse time delay estimation. The measured interference signal and the auxiliary interference signal are derived from the same source double interference structure.

6. A ranging system, characterized in that, To implement the method of claim 1, the method comprises: A swept-frequency laser is used to emit linearly frequency-modulated optical signals; An optical splitter structure is used to divide an optical signal into three paths; The time stretching module is used to perform dispersion stretching on two reference beams; The measuring interferometer, the auxiliary interferometer, and the resampling auxiliary interferometer each receive the split optical signal. The measuring interferometer and the auxiliary interferometer form a beat frequency interference signal through the time stretching module, and the resampling auxiliary interferometer provides an equal optical frequency interval interference signal. The detector module is used to receive the interference signals output by each interferometer and convert them into electrical signals.

7. A computer storage medium for storing computer programs, characterized in that, When the computer program is read by the computer, the computer executes the method of claim 1.

8. A computer, comprising a processor and a storage medium, characterized in that, When the processor reads the computer program stored in the storage medium, the computer executes the method of claim 1.

9. A computer program product, as a computer program, is characterized by: When the computer program is executed, it implements the method of claim 1.

Images