A spectral intelligent demodulation system and method with flexible bandwidth allocation function

Abstract

The present invention provides a spectral intelligent demodulation system and method with a flexible bandwidth allocation function. The system includes a signal preprocessing module, a fault detection module, a demodulation module, and a bandwidth allocation module. The signal preprocessing module is connected to the fault detection module, the fault detection module is connected to the demodulation module, and the demodulation module is connected to the bandwidth allocation module. The signal preprocessing module is used for hardware wavelet denoising, the fault detection module is used for fault detection, the demodulation module is used for spectral demodulation, and the bandwidth allocation module is used for allocating the bandwidth resources of the fiber Bragg grating sensor array. The spectral intelligent demodulation system and method with a flexible bandwidth allocation function provided by the present invention can achieve accurate demodulation of asymmetric folded spectra and can monitor and adjust the available bandwidth resources of channels according to the degree of spectral overlap.

Description

Technical Field

[0001] The present invention relates to the technical field of fiber optic sensor networks, and particularly to a spectral intelligent demodulation system and method with a flexible bandwidth allocation function. Background Art

[0002] Fiber Bragg grating (FBG) sensors have been widely used in the health monitoring of large structures such as bridges and the leakage detection of submarine oil pipelines due to their advantages of anti-electromagnetic interference and easy multiplexing. In recent years, FBG sensor networks based on wavelength division multiplexing (WDM) have important application values. With the increase in the number of FBG sensors, the capacity of fiber optic sensor networks has also been continuously increasing. Therefore, higher requirements are also put forward for the demodulation accuracy and demodulation speed of the detection system.

[0003] In recent years, to improve the multiplexing ability of WDM sensor networks, it is allowed that there is an overlap between adjacent FBG spectra. However, the overlapping spectra will increase the time and difficulty of spectral demodulation of the detection system. For this reason, domestic and foreign scholars have used intelligent optimization algorithms to achieve accurate demodulation of FBG overlapping spectra, such as simulated annealing algorithm, differential evolution algorithm, particle swarm optimization (PSO) algorithm. However, the above methods do not fully consider the influence of the FBG spectral shape. In practical applications, affected by fiber dispersion, non-uniform strain or packaging technology, the FBG spectrum is often an asymmetric spectrum. To achieve accurate demodulation of asymmetric overlapping spectra, we must clarify the influence of asymmetric parameters and asymmetric directions on spectral demodulation. On the other hand, in traditional WDM sensor networks, the occupied bandwidth of each fiber grating sensing array is fixedly allocated. When the allocated bandwidth is greater than the required bandwidth of the fiber array, it will cause waste of spectral resources on the link; when the allocated bandwidth is less than the required bandwidth of the fiber array, it will lead to serious spectral overlap of signals, and the signals are prone to distortion during transmission, which increases the difficulty of demodulating overlapping spectra. Therefore, it is very necessary to design a spectral intelligent demodulation system and method with a flexible bandwidth allocation function. Summary of the Invention

[0004] The purpose of the present invention is to provide a spectral intelligent demodulation system and method with a flexible bandwidth allocation function, which can achieve accurate demodulation of asymmetric folded spectra and can monitor and adjust the available bandwidth resources of channels according to the spectral overlap degree.

[0005] To achieve the above purpose, the present invention provides the following solutions:

[0006] A spectral intelligent demodulation system with a flexible bandwidth allocation function includes: a signal preprocessing module, a fault detection module, a demodulation module, and a bandwidth allocation module. The signal preprocessing module is connected to the fault detection module, the fault detection module is connected to the demodulation module, and the demodulation module is connected to the bandwidth allocation module;

[0007] The signal preprocessing module is used to perform hardware wavelet denoising on the noisy signal through a field-programmable gate array;

[0008] The fault detection module is used to detect faults in the fiber Bragg grating sensor array where the sampled spectral signal is located;

[0009] The demodulation module is used to perform spectral demodulation through an asymmetric overlapping spectral demodulation algorithm;

[0010] The bandwidth allocation module is used to calculate the spectral overlap degree of the sampled spectral signal and allocate the bandwidth resources of the fiber Bragg grating sensor array according to the priority of the fiber Bragg grating sensor array and the spectral overlap degree.

[0011] The present invention also provides a spectral intelligent demodulation method with a flexible bandwidth allocation function, which is applied to the spectral intelligent demodulation system with a flexible bandwidth allocation function described above, and includes the following steps:

[0012] Step 1: Preprocess the sampled spectral signal through the signal preprocessing module;

[0013] Step 2: Detect faults in the preprocessed sampled spectral signal through the fault detection module;

[0014] Step 3: Demodulate the sampled spectral signal after fault detection through the demodulation module;

[0015] Step 4: Allocate the bandwidth resources of the fiber Bragg grating sensor array through the bandwidth allocation module.

[0016] Optionally, in Step 1, preprocessing the sampled spectral signal through the signal preprocessing module specifically includes:

[0017] The signal preprocessing module collects and processes the equipment through a field-programmable gate array to perform hardware wavelet denoising on the noisy sampled spectral signal, and sends the denoised sampled spectral signal to the fault detection module.

[0018] Optionally, in Step 2, detecting faults in the preprocessed sampled spectral signal through the fault detection module specifically includes:

[0019] The fault detection module detects the sampled spectral signal, judges the faults of the fiber grating sensing array where the sampled spectral signal is located. If the sampled spectral signal disappears, it is judged that the corresponding fiber grating sensing array has a fault, and the faulty fiber grating sensing array is replaced or repaired, and the preprocessing operation is carried out again through the signal preprocessing module. If the sampled spectral signal is normal, it is judged that there is no fault, and the sampled spectral signal is sent to the demodulation module.

[0020] Optionally, in step 3, the sampled spectral signal after fault detection is demodulated by the demodulation module, specifically:

[0021] Step 301: Process the sampled spectral signal by using the CWT based on the mexh wavelet basis to achieve spectral segmentation;

[0022] Step 302: Establish an asymmetric spectral theory model and a demodulation model, and obtain the central wavelength and the asymmetry parameter of the asymmetric overlapping spectrum through the particle swarm optimization algorithm based on the asymmetric spectral theory model and the demodulation model to complete demodulation.

[0023] Optionally, in step 301, the sampled spectral signal is processed by using the CWT based on the mexh wavelet basis to achieve spectral segmentation, specifically:

[0024] Process the sampled spectral signal by using the CWT based on the mexh wavelet basis, and use the minimum value points of the transformed signal to segment the spectral signal to determine the boundaries of each spectral peak. The mexh wavelet basis function and the CWT formula are:

[0025]

[0026]

[0027] Perform continuous wavelet transform on the sampled spectral signal at different scales to obtain the wavelet coefficient matrix at multiple scales, find the maximum values of each row, and record the scales and wavelengths corresponding to each maximum value point. Among them, the scale corresponds to the row, and the wavelength corresponds to the column. If there are row maximum values continuously appearing on at least 3 scales in a certain column, the connection of the maximum values is regarded as a ridge line. Detect all the ridge lines, eliminate the ridge lines with too small energy or too low signal-to-noise ratio, determine the number of effective spectral peaks and the initial wavelength. Among them, the spectral segmentation domain and the initial central wavelength are used as the example search range and the ion initial position of the PSO algorithm to achieve spectral segmentation.

[0028] Optionally, in step 302, establish an asymmetric spectral theory model and a demodulation model, and obtain the central wavelength and the asymmetry parameter of the asymmetric overlapping spectrum through the particle swarm optimization algorithm based on the asymmetric spectral theory model and the demodulation model to complete demodulation, specifically:

[0029] The established asymmetric spectral theoretical model is as follows:

[0030]

[0031] In the formula, λ is the wavelength of the sampled spectral signal obtained by acquisition, and λ Bi is the central wavelength of the i-th FBG in the channel, and Δλ i is the 3dB bandwidth of this FBG, and α L and α R are the asymmetry parameters of the asymmetric spectrum. Among them, when the asymmetric direction is to the left, α L > 1, and α R = 1. When the asymmetric direction is to the right, α R > 1, and α L = 1. The established composite spectral model of the asymmetric overlapping spectral signal is as follows:

[0032] R′(λ) = ∑r i g i (λ - s i )

[0033] In the formula, s i is the central wavelength of the reconstructed FBG reflection spectrum. The established demodulation model is as follows:

[0034] max: G(λ) = ∫R(λ)R′(λ + τ)dτ

[0035] In the formula, R(λ) is the sampled spectral signal data of the FBG sensing network to be demodulated, R′(λ) is the spectral signal of the composite spectral model, and G(λ) is the cross-correlation function of R(λ) and R′(λ). Parameter optimization is performed through the particle swarm optimization algorithm. When G(λ) is 1, the optimal solutions of each parameter are obtained. Specifically:

[0036] Taking the segmentation domain of the spectrum as the particle search interval, randomly generate α within the range of the asymmetry parameter, set the pre-positioned central wavelength as the initial position of the particle, and initialize the velocity v - ;

[0037] Calculate the fitness function value G(λ) = ∫R(λ)R′(λ + τ)dτ of each particle, that is, the cross-correlation coefficient of R(λ) and R′(λ). The larger the cross-correlation coefficient, the closer the composite spectrum is to the sampled spectrum. Find the historical maximum value of each particle and the global maximum value of the entire particle swarm;

[0038] Update the velocity v of the particle and the central wavelength λ B of each particle, the waveform asymmetry parameter α, and determine the position of the particle after updating. Among them, the expressions for the velocity and position of the particle are:

[0039]

[0040] x p,q (t + 1) = x p,q (t) + v p,q (t + 1)

[0041] where p = 1, ..., Z, q = 1, ..., D, Z is the population size, D is the spatial dimension, t is the current iteration number, w is the weight factor, w ∈ [0, 1], c 1 and c 2 are the self - learning factor and the social - learning factor respectively, c 1 ∈ [0, 2], c 2 ∈ [0, 2], and are the individual optimal value and the global optimal value respectively, rand 1 and rand 2 are random numbers between 0 and 1. Determine whether the termination condition is reached after the update. If not, re - find the historical maximum value of each particle and the global maximum value of the entire particle swarm, and re - perform the update. If so, output the central wavelength λ B and the asymmetry parameter α L and α R .

[0042] Optionally, in step 4, the bandwidth resource of the fiber - Bragg - grating sensor array is allocated by a bandwidth allocation module. Specifically:

[0043] The bandwidth allocation module calculates the butterfly degree of the spectrum according to the overlapping area of adjacent spectra. Among them, let Δ i be the full width at half maximum of the i - th FBG reflection spectrum. If λ Bi - Δ i ≤ λ Bi-1 + Δ i-1 , then it is determined that FBG i-1 and FBG i overlap, and the spectral overlap region is [λ Bi - Δ i , λ Bi-1 + Δ i-1 . That is, the wavelength overlap region of adjacent FBGs. The overlap degree D i of the spectrum is the overlapping area between the i - th FBG reflection spectrum g i ′(λ) and the adjacent spectrum g i ′ -1 (λ), and is:

[0044]

[0045] Allocate the bandwidth resources of the fiber - Bragg - grating sensor array according to the overlap degree of the spectra.

[0046] According to the specific embodiments provided by the present invention, the following technical effects are disclosed by the present invention: The spectral intelligent demodulation system and method with flexible bandwidth allocation function provided by the present invention, the system includes a signal preprocessing module, a fault detection module, a demodulation module and a bandwidth allocation module. Among them, the signal preprocessing module is used to perform hardware wavelet denoising on the noisy signal through a field programmable gate array, the fault detection module is used to detect faults in the fiber Bragg grating sensor array where the sampled spectral signal is located, the demodulation module is used to perform spectral demodulation through an asymmetric overlapping spectral demodulation algorithm, and the bandwidth allocation module is used to calculate the spectral overlapping degree of the sampled spectral signal and allocate the bandwidth resources of the fiber Bragg grating sensor array according to the priority and spectral overlapping degree of the fiber Bragg grating sensor array; the method includes preprocessing the sampled spectral signal through the signal preprocessing module, detecting faults in the preprocessed sampled spectral signal through the fault detection module, demodulating the sampled spectral signal after fault detection through the demodulation module, and allocating the bandwidth resources of the fiber Bragg grating sensor array through the bandwidth allocation module; the method can ensure the accurate demodulation of symmetric overlapping spectra and can also achieve the accurate demodulation of asymmetric overlapping spectra, can effectively reduce the adverse effects brought by asymmetry and overlap, improve the multiplexing ability of the FBG sensor network. During the demodulation process of the method, the bandwidth resources of the array are adjusted according to the priority of the fiber Bragg grating sensing array and the spectral overlapping degree, reducing the spectral overlapping degree of adjacent FBGs and reducing the demodulation difficulty, which can provide a reference for realizing the bandwidth resource optimization of elastic optical networks. The method can quickly locate through the fault detection module to detect faults in the fiber Bragg grating sensor array where the spectral signal is located, and can switch to a standby fiber Bragg grating sensor array to ensure the normal operation of the system. Description of the Drawings

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0048] Figure 1 It is a schematic diagram of the FBG sensing network structure;

[0049] Figure 2 It is a schematic diagram of the flow of the spectral intelligent demodulation method with flexible bandwidth allocation function in the embodiment of the present invention;

[0050] Figure 3 It is a schematic diagram of asymmetric overlapping FBG spectra and CWT signals in various cases;

[0051] Figure 4Schematic diagram of the extracted wavelet ridge line;

[0052] Figure 5 Schematic diagram of the parameter optimization process by the particle swarm optimization algorithm. Specific implementation manners

[0053] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0054] The object of the present invention is to provide a spectral intelligent demodulation system and method with a flexible bandwidth allocation function, which can achieve accurate demodulation of asymmetric folded spectra and can monitor and adjust the available bandwidth resources of channels according to the spectral overlap degree.

[0055] To make the above objects, features and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners.

[0056] As Figure 1 shown, the FBG sensor network has multiple fiber Bragg grating sensor arrays, but the priorities of each fiber Bragg grating sensor array are different, and there are differences in the spectral overlap degree. The FBG reflected spectral signal is generally sent to a spectral analyzer for sampling through a 3dB coupler, and the sampled data is sent to a computer for data analysis after filtering and denoising.

[0057] As Figure 2 shown, the spectral intelligent demodulation system with a flexible bandwidth allocation function provided by the embodiment of the present invention includes: a signal preprocessing module, a fault detection module, a demodulation module, and a bandwidth allocation module. The signal preprocessing module is connected to the fault detection module, the fault detection module is connected to the demodulation module, and the demodulation module is connected to the bandwidth allocation module;

[0058] The signal preprocessing module is used to perform hardware wavelet denoising on the noisy signal through a field programmable gate array;

[0059] The fault detection module is used to detect faults in the fiber Bragg grating sensor array where the sampled spectral signal is located;

[0060] The demodulation module is used to perform spectral demodulation through an asymmetric overlapping spectral demodulation algorithm;

[0061] The bandwidth allocation module is used to calculate the spectral overlap degree of the sampled spectral signal and allocate the bandwidth resources of the fiber Bragg grating sensor array according to the priority of the fiber Bragg grating sensor array and the spectral overlap degree.

[0062] As shown Figure 2 in the figure, the present invention also provides a spectral intelligent demodulation method with a flexible bandwidth allocation function, which is applied to the above spectral intelligent demodulation system with a flexible bandwidth allocation function, and includes the following steps:

[0063] Step 1: Preprocess the sampled spectral signal through a signal preprocessing module;

[0064] Step 2: Perform fault detection on the preprocessed sampled spectral signal through a fault detection module;

[0065] Step 3: Demodulate the sampled spectral signal after fault detection through a demodulation module;

[0066] Step 4: Allocate the bandwidth resources of the fiber Bragg grating sensor array through a bandwidth allocation module.

[0067] In Step 1, the preprocessing of the sampled spectral signal through the signal preprocessing module is specifically as follows:

[0068] The signal preprocessing module performs hardware wavelet denoising on the noisy sampled spectral signal through a field programmable gate array acquisition and processing device, and sends the denoised sampled spectral signal to the fault detection module.

[0069] In Step 2, the fault detection of the preprocessed sampled spectral signal through the fault detection module is specifically as follows:

[0070] The fault detection module detects the sampled spectral signal, judges the fault of the fiber Bragg grating sensing array where the sampled spectral signal is located. If the sampled spectral signal disappears, it is judged that the corresponding fiber Bragg grating sensing array has a fault, replaces or repairs the faulty fiber Bragg grating sensing array, and performs the preprocessing operation again through the signal preprocessing module. If the sampled spectral signal is normal, it is judged that there is no fault, and the sampled spectral signal is sent to the demodulation module.

[0071] In Step 3, the demodulation of the sampled spectral signal after fault detection through the demodulation module is specifically as follows:

[0072] Step 301: Process the sampled spectral signal using continuous wavelet transform (CWT) based on the mexh wavelet basis to achieve spectral segmentation;

[0073] Step 302: Establish an asymmetric spectral theory model and a demodulation model, and obtain the central wavelength and asymmetry parameter of the asymmetric overlapping spectrum through a particle swarm optimization algorithm based on the asymmetric spectral theory model and the demodulation model to complete demodulation.

[0074] In step 301, the sampled spectral signal is processed using CWT based on the mexh wavelet basis to achieve spectral segmentation. Specifically:

[0075] As Figure 3 shown, the sampled spectral signal is processed using CWT based on the mexh wavelet basis. The minimum points of the transformed signal are used to segment the spectral signal to determine the boundaries of each spectral peak. The mexh wavelet basis function and the CWT formula are:

[0076]

[0077]

[0078] As Figure 4 shown, it shows how to use the wavelet ridge to determine the number of overlapping spectra and achieve the pre - positioning of the central wavelength;

[0079] The sampled spectral signal is subjected to continuous wavelet transform at different scales to obtain the wavelet coefficient matrix at multiple scales. The maximum values of each row are obtained, and the scales and wavelengths corresponding to each maximum value point are recorded. Among them, the scale corresponds to the row, and the wavelength corresponds to the column. If there are row maximum values continuously appearing at least at 3 scales in a certain column, the connection of the maximum values is regarded as a ridge line. All ridge lines are detected, and the ridge lines with too small energy or too low signal - to - noise ratio are removed to determine the number of effective spectral peaks and the initial wavelength. Among them, the spectral segmentation method based on CWT uses the spectral segmentation domain and the initial central wavelength as the particle search interval and the initial particle position of the PSO algorithm to achieve spectral segmentation.

[0080] In step 302, an asymmetric spectral theoretical model and a demodulation model are established. Based on the asymmetric spectral theoretical model and the demodulation model, the central wavelength and the asymmetry parameter of the asymmetric overlapping spectrum are obtained through the particle swarm optimization algorithm to complete the demodulation. Specifically:

[0081] The intelligent optimization algorithm is used to demodulate the asymmetric overlapping spectrum. First, it is transformed into a function optimization problem. The asymmetric spectral theoretical model is established as:

[0082]

[0083] In the formula, λ is the wavelength of the sampled spectral signal collected, λ Bi is the central wavelength of the i - th FBG in the channel, Δλ i is the 3dB bandwidth of this FBG, α L and α R are the asymmetry parameters of the asymmetric spectrum. Among them, when the asymmetric direction is to the left, α L > 1, α R = 1. When the asymmetric direction is to the right, α R > 1, α L= 1. Since there are multiple FBG sensors multiplexed in the fiber Bragg grating sensing array, the synthetic spectral model for the asymmetric overlapping spectral signals is established as follows:

[0084] R′(λ) = ∑r i g i (λ - s i )

[0085] where s i is the central wavelength of the reconstructed FBG reflection spectrum. The demodulation model is established as:

[0086] max: G(λ) = ∫R(λ)R′(λ + τ)dτ

[0087] where R(λ) is the sampled spectral signal data of the FBG sensing network to be demodulated, R′(λ) is the spectral signal of the synthetic spectral model, and G(λ) is the cross - correlation function of R(λ) and R′(λ). Parameter optimization is carried out through the particle swarm optimization algorithm. When G(λ) is 1, the optimal solutions of each parameter are obtained. The specific steps are as Figure 5 shown, specifically:

[0088] Taking the spectral segmentation domain as the particle search interval, randomly generate α within the range of the asymmetric parameter, set the predetermined central wavelength as the initial position of the particle, and initialize the velocity v - ;

[0089] Calculate the fitness function value of each particle G(λ) = ∫R(λ)R′(λ + τ)dτ, that is, the cross - correlation coefficient of R and R′. The larger the cross - correlation coefficient, the closer the synthetic spectrum is to the sampled spectrum. Find the historical maximum value of each particle and the global maximum value of the entire particle swarm;

[0090] Update the velocity v of the particle and the central wavelength λ of each particle B , the waveform asymmetry parameter α, and determine the position of the particle after updating. Among them, the expressions for the velocity and position of the particle are:

[0091]

[0092] x p,q (t + 1) = x p,q (t) + v p,q (t + 1)

[0093] where p = 1,..., Z, q = 1,..., D, Z is the population size, D is the space dimension, t is the current iteration number, w is the weight factor, w ∈ [0, 1], c 1 and c 2 are the self - learning factor and the social - learning factor respectively, c 1 ∈ [0, 2], c 2∈[0, 2], and are the individual optimal value and the global optimal value respectively, rand 1 and rand 2 are random numbers between 0 and 1. It is judged whether the termination condition is reached after the update. If not, the historical maximum value of each particle and the global maximum value of the entire particle swarm are searched again and updated again. If so, the particle swarm algorithm outputs the optimal solution to obtain the central wavelength λ B and the asymmetry parameter α L and α R .

[0094] In step 4, the bandwidth resources of the fiber Bragg grating sensor array are allocated by the bandwidth allocation module. Specifically:

[0095] The bandwidth allocation module calculates the degree of butterfly formation of the spectrum according to the overlapping area of adjacent spectra. Among them, let Δ i be the full width at half maximum of the reflection spectrum of the i-th FBG. If λ Bi -Δ i ≤λ Bi-1 +Δ i-1 , then it is judged that FBG i-1 overlaps with FBG i , and the spectral overlapping area is [λ Bi -Δ i , λ Bi-1 +Δ i-1 , that is, the wavelength overlapping area of adjacent FBGs. The overlapping degree D i of the spectrum is the overlapping area between the reflection spectrum g′ i (λ) of the i-th FBG and the adjacent spectrum g′ i-1 (λ), and is:

[0096]

[0097] The bandwidth resources of the fiber Bragg grating sensor array are allocated according to the overlapping degree of the spectrum. Among them, the overlapping degree of the reflection spectra of the entire fiber Bragg grating sensing array is defined as the average overlapping degree, and the influence of the asymmetry parameter and the asymmetry direction on the overlapping degree of the spectrum is taken into account: the larger the asymmetry parameter, the larger Δ i , the easier the spectrum is to overlap; if the asymmetry direction of the spectrum is towards the overlapping area, the overlapping degree of the spectrum increases. When the priority of a certain fiber Bragg grating sensor array is high and the corresponding spectral overlapping degree is too large, the bandwidth of the fiber Bragg grating sensor array can be appropriately increased to reduce the spectral overlapping degree and the demodulation difficulty of the asymmetric overlapping spectrum.

[0098] The spectral intelligent demodulation system and method with flexible bandwidth allocation function provided by the present invention. The system includes a signal preprocessing module, a fault detection module, a demodulation module, and a bandwidth allocation module. Among them, the signal preprocessing module is used to perform hardware wavelet denoising on the noisy signal through a field-programmable gate array. The fault detection module is used to detect faults in the fiber Bragg grating sensor array where the sampled spectral signal is located. The demodulation module is used to perform spectral demodulation through an asymmetric overlapping spectral demodulation algorithm. The bandwidth allocation module is used to calculate the spectral overlapping degree of the sampled spectral signal and allocate the bandwidth resources of the fiber Bragg grating sensor array according to the priority and spectral overlapping degree of the fiber Bragg grating sensor array. The method includes preprocessing the sampled spectral signal through the signal preprocessing module, detecting faults in the preprocessed sampled spectral signal through the fault detection module, demodulating the sampled spectral signal after fault detection through the demodulation module, and allocating the bandwidth resources of the fiber Bragg grating sensor array through the bandwidth allocation module. This method can ensure the accurate demodulation of symmetric overlapping spectra and also achieve the accurate demodulation of asymmetric overlapping spectra. It can effectively reduce the adverse effects brought by asymmetry and overlap, improve the multiplexing ability of the FBG sensor network. During the demodulation process, this method adjusts the bandwidth resources of the array according to the priority of the fiber Bragg grating sensing array and the spectral overlapping degree, reduces the overlapping degree of adjacent FBG spectra, and reduces the demodulation difficulty. It can provide a reference for realizing the optimization of bandwidth resources in elastic optical networks. This method can quickly locate faults by detecting faults in the fiber Bragg grating sensor array where the spectral signal is located, and can switch to a standby fiber Bragg grating sensor array to ensure the normal operation of the system.

[0099] In this article, specific examples are used to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention. At the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A spectral intelligent demodulation method with flexible bandwidth allocation function, applied to a spectral intelligent demodulation system with flexible bandwidth allocation function, characterized in that, the system includes: a signal preprocessing module, a fault detection module, a demodulation module and a bandwidth allocation module. The signal preprocessing module is connected to the fault detection module, the fault detection module is connected to the demodulation module, and the demodulation module is connected to the bandwidth allocation module; the signal preprocessing module is used to perform hardware wavelet denoising on the noisy signal through a field programmable gate array; the fault detection module is used to detect faults in the fiber Bragg grating sensor array where the sampled spectral signal is located; the demodulation module is used to perform spectral demodulation through an asymmetric overlapping spectrum demodulation algorithm; the bandwidth allocation module is used to calculate the spectral overlap degree of the sampled spectral signal and allocate bandwidth resources for the fiber Bragg grating sensor array according to the priority of the fiber Bragg grating sensor array and the spectral overlap degree; the method includes the following steps: Step 1: Preprocess the sampled spectral signal through the signal preprocessing module; Step 2: Detect faults in the preprocessed sampled spectral signal through the fault detection module; Step 3: Demodulate the sampled spectral signal after fault detection through the demodulation module; Step 4: Allocate bandwidth resources for the fiber Bragg grating sensor array through the bandwidth allocation module; In Step 3, demodulating the sampled spectral signal after fault detection through the demodulation module specifically includes: Step 301: Process the sampled spectral signal using continuous wavelet transform (CWT) based on the mexh wavelet basis to achieve spectral segmentation; Step 302: Establish an asymmetric spectral theory model and a demodulation model, and obtain the central wavelength and asymmetry parameter of the asymmetric overlapping spectrum through a particle swarm optimization algorithm based on the asymmetric spectral theory model and the demodulation model to complete demodulation; In Step 302, establishing an asymmetric spectral theory model and a demodulation model, and obtaining the central wavelength and asymmetry parameter of the asymmetric overlapping spectrum through a particle swarm optimization algorithm based on the asymmetric spectral theory model and the demodulation model to complete demodulation specifically includes: The established asymmetric spectral theory model is: where λ is the wavelength of the sampled spectral signal obtained by acquisition, λ Bi is the central wavelength of the i-th FBG in the channel, and Δλ i is the 3dB bandwidth of the FBG, and α L and α R are the asymmetry parameters of the asymmetric spectrum. Among them, when the asymmetric direction is to the left, α L > 1, α R = 1. When the asymmetric direction is to the right, α R > 1, α L = 1. The synthetic spectral model of the asymmetric overlapping spectral signal is established as: R′(λ) = ∑r i g′ i (λ - s i ) where s i is the central wavelength of the reconstructed FBG reflection spectrum, and the demodulation model is established as follows: max: G(λ) = ∫R(λ)R′(λ + τ)dτ where R(λ) is the sampled spectral signal data of the FBG sensing network to be demodulated, R′(λ) is the spectral signal of the synthetic spectral model, G(λ) is the cross-correlation function of R(λ) and R′(λ), and parameter optimization is performed through a particle swarm optimization algorithm. When G(λ) is 1, the optimal solutions of each parameter are obtained. Specifically: Using the segmented domain of the spectrum as the particle search interval, randomly generate α within the range of the asymmetry parameter, set the pre-located central wavelength as the initial position of the particle, and initialize the velocity v - ; Calculate the fitness function value G(λ) = ∫R(λ)R′(λ + τ)dτ of each particle, that is, the cross-correlation coefficient of R(λ) and R′(λ). The larger the cross-correlation coefficient, the closer the synthetic spectrum is to the sampled spectrum. Find the historical maximum value of each particle and the global maximum value of the entire particle swarm; Update the velocity v of the particles and the central wavelength λ of each particle B , the waveform asymmetry parameter α, and determine the updated position of the particles. Among them, the expressions for the velocity and position of the particles are as follows: x p,q (t + 1) = x p,q (t) + v p,q (t + 1) where p = 1, ..., Z, q = 1, ..., D, Z is the population size, D is the spatial dimension, t is the current iteration number, w is the weight factor, w ∈ [0, 1], c 1 and c 2 are the self-learning factor and the social learning factor respectively, c 1 ∈ [0, 2], c 2 ∈ [0, 2], and are the individual optimal value and the global optimal value respectively, rand 1 and rand 2 are random numbers between 0 and 1. It is judged whether the termination condition is reached after the update. If not, the historical maximum value of each particle and the global maximum value of the entire particle swarm are searched again and updated again. If so, the central wavelength λ B and the asymmetry parameter α L and α R ; In Step 4, allocating bandwidth resources for the fiber Bragg grating sensor array through the bandwidth allocation module specifically includes: The bandwidth allocation module calculates the overlapping degree of spectra based on the overlapping area of adjacent spectra. Among them, let Δ i be the full width at half maximum of the reflection spectrum of the i-th FBG. If λ Bi -Δ i ≤λ Bi-1 +Δ i-1 , it is determined that FBG i-1 overlaps with FBG i , and the spectral overlapping region is [λ Bi -Δ i , λ Bi-1 +Δ i-1 , that is, the wavelength overlapping region of adjacent FBGs. The overlapping degree D i of the spectra is the overlapping area between the reflection spectrum g′ i (λ) of the i-th FBG and the adjacent spectrum g′ i-1 (λ), and is: Allocating bandwidth resources for the fiber Bragg grating sensor array according to the spectral overlap degree.

2. The spectral intelligent demodulation method with flexible bandwidth allocation function according to claim 1, It is characterized in that In step 1, the sampled spectral signal is preprocessed by a signal preprocessing module, specifically: The signal preprocessing module performs hardware wavelet denoising on the noisy sampled spectral signal through a field editable gate array acquisition and processing device, and sends the denoised sampled spectral signal to the fault detection module.

3. The spectrum intelligent demodulation method with flexible bandwidth allocation function according to claim 2, It is characterized in that In step 2, the fault detection module performs fault detection on the preprocessed sampled spectral signal, specifically: The fault detection module detects the sampling spectrum signal and makes a fault judgment on the fiber grating sensor array where the sampling spectrum signal is located. If the sampling spectrum signal disappears, it is judged that the corresponding fiber grating sensor array is faulty, and the faulty fiber grating sensor array is replaced or repaired, and the signal preprocessing module is used to perform preprocessing operations again. If the sampling spectrum signal is normal, it is judged that no fault occurs, and the sampling spectrum signal is sent to the demodulation module.

4. The spectrum intelligent demodulation method with flexible bandwidth allocation function according to claim 3, It is characterized in that In step 301, the sampled spectral signal is processed using CWT based on mexh wavelet basis to achieve spectral segmentation, specifically: The sampled spectral signal is processed by using CWT based on mexh wavelet basis, and the minimum value point of the transformed signal is used to segment the spectral signal to determine the boundary of each spectral peak. The sampled spectral signal is subjected to continuous wavelet transform at different scales to obtain the wavelet coefficient matrix at multiple scales, the maximum value of each row is obtained, and the scale and wavelength corresponding to each maximum point are recorded, wherein the scale corresponds to the row, and the wavelength corresponds to the column. If there are row maxima on a certain column that appear continuously at at least three scales, the connecting line of the maxima is regarded as a ridge line, all ridge lines are detected, and ridge lines with too small energy or too low signal-to-noise ratio are eliminated to determine the number of effective spectral peaks and the initial wavelength, wherein the spectral segmentation domain and the initial central wavelength are used as examples of the PSO algorithm to search for the extraction and the initial position of the ion to achieve spectral segmentation.

Images