Multi-node edge induction sampling system and method based on environmental monitoring

Abstract

The invention discloses a multi-node edge induction sampling system and method based on environmental monitoring, and the method comprises the steps: firstly employing a radio frequency signal transmitted by a central transmitting unit as a stable signal source, and providing a stable and consistent signal source for a subsequent frequency mixing process; the method comprises the following steps: mixing a noise signal acquired by using a portable sensor anytime and anywhere with a reference radio frequency signal to obtain a frequency mixing signal; extracting a noise signal from the frequency mixing signal; the noise level is obtained through analysis according to the intensity or frequency spectrum characteristics of the noise signals, and when the noise level exceeds a set threshold value, the system sends out a warning signal in time to remind workers to pay attention to noise hazards and prevent the workers from being exposed in a harmful noise environment for a long time. By providing real-time noise level evaluation and warning, the system not only can monitor the noise environment, but also can warn the safety of the staff to ensure that the staff work in a healthy range.

Description

Technical Field

[0001] The present invention relates to the field of noise sensing sampling, and in particular to a multi-node edge sensing sampling system and method based on environmental monitoring. Background Art

[0002] Productive noise is generated by the impact, friction, and rotation of machinery during the production process, which has a significant impact on human health, especially on the auditory and nervous systems; long-term exposure to high-noise environments may cause hearing loss, neurasthenia, cardiovascular disease and other occupational diseases; overall noise source control, strict industrial noise management, and promotion of advanced technologies are important measures to strengthen noise monitoring. In industrial enterprises, taking vibration and noise reduction measures, strengthening noise source management, and achieving production noise pollution control are important technical advancement goals;

[0003] The first condition for achieving this technical recommendation is to collect and analyze the noise level generated during the production process;

[0004] The traditional way to collect and analyze noise levels in the production process is for each employee to carry a huge measurement sampler, and then use the measurement sampler to collect the noise level of each employee's fixed work point at a fixed time (that is, the time when the employee works at the permanent work point);

[0005] However, the huge size of the measurement sampler makes it inconvenient for employees to work. At the same time, because it can only collect noise levels at fixed locations and fixed times, it cannot conveniently collect noise levels at various locations and times on site, and thus cannot directly reflect the actual noise level on site. Summary of the invention

[0006] The purpose of the present invention is to provide a multi-node edge sensing sampling system and method based on environmental monitoring, which solves the above-mentioned technical problems pointed out in the prior art.

[0007] The present invention provides a multi-node edge sensing sampling system based on environmental monitoring, including a central transmitting unit, a data receiving module, a mixing module, a processing and screening module, and an analysis and warning module;

[0008] The central transmitting unit is used to transmit radio frequency signals in real time in the current noise place; the data receiving module is used to receive radio frequency signals sent by the central transmitting unit in real time in the current noise place; the data receiving module is a sensor with a signal receiver function worn by each individual; the mixed signal is used to obtain a mixed signal based on the mixing of the radio frequency signal and the sound signal generated in the current noise place; the processing and screening module is used to perform Fourier transform processing and screening processing on the mixed signal to obtain a noise signal; the analysis and warning module is used to analyze the noise signal to obtain a noise level; the noise level is stored and displayed, and a warning signal is issued to the current staff based on the noise level.

[0009] Correspondingly, the present invention also proposes a multi-node edge sensing sampling method based on environmental monitoring, including the following operating steps: when the staff is in the current noisy place, the portable sensor worn by each staff member receives the radio frequency signal sent from the central transmitting unit in real time; based on the mixing of the radio frequency signal with the sound signal generated by the current noisy place, a mixed signal is obtained; the mixed signal is Fourier transformed and filtered to obtain a noise signal; based on the noise signal, the noise level is analyzed to obtain the noise level; the noise level is stored and displayed, and a warning signal is issued to the current staff member based on the noise level.

[0010] Compared with the prior art, the embodiments of the present invention have at least the following technical advantages:

[0011] By analyzing the above-mentioned multi-node edge sensing sampling system and method based on environmental monitoring provided by the present invention, it can be known that in specific applications, the radio frequency signal emitted by the central transmitting unit is first used as a stable signal source to provide a stable and consistent signal source for the subsequent mixing process, ensuring that the system can accurately extract and analyze the noise signal; by mixing the noise signal collected by the portable sensor anytime and anywhere with the reference radio frequency signal, the system can extract and process noise signals of different frequencies, providing a data basis for subsequent noise analysis; further, components related to productive noise are extracted from the mixed signal, ensuring that the system can accurately identify noise signals that have an impact on human health from complex mixed signals, laying the foundation for subsequent noise level evaluation; the noise level is analyzed according to the intensity or spectrum characteristics of the noise signal. When the noise level exceeds the set threshold, the system promptly issues a warning signal to remind the staff to pay attention to the noise hazard and avoid long-term exposure to harmful noise environment; by providing real-time noise level evaluation and warning, the system can not only monitor the noise environment, but also issue safety warnings to the staff to ensure that they work within a healthy range. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A schematic diagram of the overall architecture of a multi-node edge sensing sampling system based on environmental monitoring provided in Example 1;

[0013] Figure 2 A schematic diagram of the operation steps of a multi-node edge sensing sampling method based on environmental monitoring provided in Example 2.

[0014] Figure numerals: central transmitting unit 10, data receiving module 20, mixing module 30, processing and screening module 40, analysis and warning module 50, transformation processing module 41, feature extraction and screening module 42, matching module 43, screening and analysis module 44, initial screening module 421, re-screening module 422, merging module 423, time window division module 4221, frequency component acquisition module 4222, first calculation module 4223, second calculation module 4224, judgment module 4225, initialization module 42241, matrix construction module 42242, third calculation module 42243, fourth calculation module 42244, fifth calculation module 42245, sixth calculation module 42246, seventh calculation module 42247, judgment output module 42248. DETAILED DESCRIPTION

[0015] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0016] The present invention is further described in detail below through specific embodiments in conjunction with the accompanying drawings.

[0017] Embodiment 1

[0018] like Figure 1 As shown, the first embodiment of the present invention provides a multi-node edge sensing sampling system based on environmental monitoring, including a central transmitting unit 10, a data receiving module 20, a mixing module 30, a processing and screening module 40, and an analysis and warning module 50;

[0019] The central transmitting unit 10 is used to transmit a radio frequency signal in real time in a current noise place; the data receiving module 20 is used to receive a radio frequency signal sent by the central transmitting unit in real time in the current noise place; the mixing signal 30 is used to obtain a mixing signal based on mixing the radio frequency signal with a sound signal generated in the current noise place; the processing and screening module 40 is used to perform Fourier transform processing and screening processing on the mixing signal to obtain a noise signal; the analysis and warning module 50 is used to analyze the noise signal to obtain a noise level; the noise level is stored and displayed, and a warning signal is issued to the current staff based on the noise level.

[0020] The above-mentioned sampling system adopts a central setting method to send a stable radio frequency signal or induction signal to the surrounding space, and then each person wears a sensor with a signal receiver function (i.e., a data receiving module 20), which is used to sense and receive the signal sent by the central transmitting unit 10, thereby ensuring that the central multi-node induction sampling system can identify the signals at different positions of multiple nodes (each node is equivalent to an individual worker), so as to calculate the real-time noise reception measurement of the current node; after receiving the noise exposure of individual workers at multiple node positions on a large scale, the noise (productive noise) reception measurement of different workers in the same place, i.e., different nodes, can be compared, thereby providing a technical basis for the subsequent objective evaluation of the noise (productive noise) reception measurement of the node throughout the year or a certain time period.

[0021] Since multiple central transmitting units 10 are set up in the venue, individual staff only need to walk to different locations to connect to the central transmitting unit 10 nearby, and the central transmitting unit 10 will complete the noise level collection, so there is no need for everyone to carry a huge measurement sampler.

[0022] Preferably, the processing and screening module 40 includes a transformation processing module 41, a feature extraction and screening module 42, a matching module 43 and a screening analysis module 44;

[0023] The transform processing module 41 is used to perform Fourier transform processing based on the mixed signal to obtain spectrum information S y (f);

[0024] The feature extraction and screening module 42 is used to perform a time domain feature extraction operation on the mixing signal of the observation period T to obtain time domain feature information; based on the time domain feature information, the mixing signal is screened to obtain target spectrum information S y (f');

[0025] The matching module 43 is used to match the target spectrum information Sy (f') Acquire multiple spectrum ranges; Based on the preset mechanical noise spectrum S mech (f) and the mechanical noise spectrum S class (f) Combine the spectrum range to obtain the typical mechanical noise spectrum Score mech Compared with the typical mechanical noise spectrum Score class ;

[0026] The screening and analysis module 44 is used to analyze the typical mechanical noise spectrum Score based on the mech Compared with the typical mechanical noise spectrum Score class And the time domain characteristic information is screened and analyzed to obtain a noise signal;

[0027] Preferably, the time domain feature information includes the root mean square RMS value RMS(y(t)) and the autocorrelation function R y (τ).

[0028] Preferably, the screening and analysis module 44 is specifically used to determine the typical mechanical noise spectrum Score mech Is the value greater than the typical mechanical noise spectrum Score? class If yes, then determine the target spectrum information S corresponding to the spectrum range y (f') is the mechanical noise signal to be determined; if not, the typical mechanical noise spectrum Score is determined mech Is the value less than the typical mechanical noise spectrum Score? class If yes, then determine the target spectrum information S corresponding to the spectrum range y (f') is the mechanical noise signal to be determined;

[0029] Preferably, the feature extraction and screening module 42 includes a primary screening module 421, a secondary screening module 422 and a merging module 423;

[0030] The initial screening module 421 is used to perform initial screening on the mixed signal based on the time domain feature information to obtain the first target spectrum information S y (f')1 and the mixed signal to be screened;

[0031] The re-screening module 422 is used to perform a re-screening operation based on the time dimension of the mixed signal to be screened and the frequency component of the mixed signal to be screened to obtain the second target spectrum information S y (f')2;

[0032] The merging module 423 is used to combine the first target spectrum information S y (f')1 and the second target spectrum information Sy (f')2 Get the target spectrum information S y (f').

[0033] Preferably, the re-screening module 422 includes a time window division module 4221, a frequency component acquisition module 4222, a first calculation module 4223, a second calculation module 4224 and a judgment module 4225;

[0034] The time window division module 4221 is used to divide the mixed frequency signal to be screened according to the time sequence corresponding to the mixed frequency signal to be screened to obtain a plurality of time windows i={i1, i2, i3, ...iN}; wherein iN is the Nth time window;

[0035] The frequency component acquisition module 4222 is used to obtain the frequency component j={j1, j2, j3, ...jK} corresponding to each time window i by short-time Fourier transform processing; wherein jK is the frequency component of the Kth time window;

[0036] The first calculation module 4223 is used to calculate the time-related index TC(i) based on the time-frequency matrix M;

[0037] The second calculation module 4224 is used to perform iterative calculation based on the time-frequency matrix M and in combination with the time scale and frequency scale of the time-frequency matrix M to obtain the spectrum stability index SF; calculate the comprehensive scoring function Score(i) based on the spectrum stability index SF and the time-related index TC(i);

[0038] The judgment module 4225 is used to judge whether the comprehensive scoring function Score(i) is greater than or equal to a preset comprehensive scoring function threshold. If so, it is determined that the mixed signal to be screened corresponding to the time window is the second target spectrum information S y (f')2.

[0039] Preferably, the second calculation module 4224 includes an initialization module 42241, a matrix construction module 42242, a third calculation module 42243, a fourth calculation module 42244, a fifth calculation module 42245, a sixth calculation module 42246, a seventh calculation module 42247 and a judgment output module 42248;

[0040] The initialization module 42241 is used to construct an initial spectrum stability matrix SSM(0) based on the time-frequency matrix M; calculate a density index D(i) based on the time-frequency matrix M; and initialize iteration parameters;

[0041] The matrix construction module 42242 is used for the iteration parameters including the smoothing coefficient α, the initial spectrum stability index SF(i), the time scale parameter τt, the frequency scale parameter τf, the weight coefficient β1, the weight coefficient β2, the weight coefficient β3, the learning rate γ, the iteration counter and the maximum number of iterations threshold; the iteration number of the iteration counter is initially 0;

[0042] The third calculation module 42243 is used to construct a time domain similarity matrix U based on the time window and the time scale parameter τt; and to construct a frequency domain similarity matrix V based on the frequency component and the frequency scale parameter τf;

[0043] The fourth calculation module 42244 is used to iteratively update the spectrum stability matrix: add 1 to the iteration number of the iteration counter to obtain the current iteration number; and calculate based on the smoothing coefficient α, the initial spectrum stability matrix SSM(0), the time domain similarity matrix U and the frequency domain similarity matrix V to obtain the updated spectrum stability matrix SSM(z+1);

[0044] The fifth calculation module 42245 is used to calculate and obtain a target eigenvector C(i) by solving an eigenvalue equation based on the updated spectrum stability matrix SSM(z+1);

[0045] The sixth calculation module 42246 is used to calculate the updated spectrum stability index SF based on the target feature vector C(i), the density index D(i) and the initial spectrum stability index SF(i). new (i);

[0046] The seventh calculation module 42247 is used to calculate the target convergence threshold under the current number of iterations based on the current number of iterations combined with the preset initial convergence threshold; and based on the updated spectrum stability index SF new (i) and the updated spectrum stability matrix SSM(z+1) to calculate the comprehensive convergence value TC;

[0047] The judgment output module 42248 is used to judge whether the comprehensive convergence value is less than or equal to the target convergence threshold. If so, output the updated spectrum stability index SF new (i) is the target spectrum stability index SF; if not, determining whether the current number of iterations is greater than or equal to a preset maximum number of iterations threshold, and if so, outputting the updated spectrum stability index SF new(i) is the target spectrum stability index SF; if not, then the smoothing coefficient α is updated based on the updated spectrum stability matrix SSM(z+1) to obtain the updated smoothing coefficient α'; the updated smoothing coefficient α' is used as the smoothing coefficient α, and the above-mentioned iterative update of the spectrum stability matrix operation is returned and re-executed until the target spectrum stability index SF is output.

[0048] Preferably, the seventh calculation module 42247 is specifically used to calculate the relative change rate RC of the spectrum stability matrix based on the updated spectrum stability matrix SSM(z+1); based on the updated spectrum stability index SF new (i) Calculating the relative change rate SC of the stability index;

[0049] The comprehensive convergence value TC is calculated based on the relative change rate RC of the spectrum stability matrix and the relative change rate SC of the stability index.

[0050] Preferably, the judgment output module 42248 is specifically used to calculate the updated smoothing coefficient α' based on the relative change rate RC of the spectrum stability matrix.

[0051] Preferably, the updated smoothing coefficient α' is calculated as follows:

[0052] α'=α×(1+γ×RC);

[0053] In the formula, γ is the learning rate.

[0054] In summary, the multi-node edge sensing sampling system based on environmental monitoring proposed in the embodiment of the present application first transmits a radio frequency signal through the central transmitting unit 10 as a stable signal source to provide a reference benchmark for subsequent noise analysis; then, the data receiving module 20 receives the radio frequency signal from the central transmitting unit; then, the mixing module 30 generates a mixed signal by mixing the received radio frequency signal with the sound signal in the noisy place, laying the foundation for subsequent noise extraction and analysis;

[0055] In the processing and screening stage, the transformation processing module 41 in the processing and screening module 40 performs Fourier transform on the mixed signal to obtain spectrum information; the feature extraction and screening module 42 performs time domain analysis on the mixed signal based on time domain feature extraction, extracts time domain feature information, and screens to obtain target spectrum information; the matching module 43 obtains multiple spectrum ranges according to the target spectrum information, and matches them with the preset mechanical noise spectrum and quasi-mechanical noise spectrum to determine the typical mechanical noise and quasi-mechanical noise spectrum; the screening analysis module 44 uses these spectrum information and time domain feature information for further analysis to accurately identify the noise signal;

[0056] In the re-screening process, the re-screening module 422 first divides the mixed signal into multiple time windows according to the time series through the time window division module 4221, and performs short-time Fourier transform through the frequency component acquisition module 4222 to obtain the frequency component of each time window; then, the first calculation module 4223 calculates the time-related index based on the time-frequency matrix M, and the second calculation module 4224 performs iterative calculation based on the time and frequency scales of the time-frequency matrix to obtain the spectrum stability index SF, and calculates the comprehensive scoring function based on these indicators; the judgment module 4225 judges whether it exceeds the set threshold according to the comprehensive scoring function, and if it meets the conditions, the second target spectrum information is screened out;

[0057] Further, the second calculation module 4224 constructs an initial spectrum stability matrix SSM(0) through the initialization module 42241, and calculates the density index; the matrix construction module 42242 sets the iteration parameters, including the smoothing coefficient, the time scale, the frequency scale, etc., and starts the iterative calculation. During the iteration process, the spectrum stability matrix is ​​continuously updated, and the time domain similarity matrix U and the frequency domain similarity matrix V are updated respectively through the third calculation module 42243 and the fourth calculation module 42244; the fifth calculation module 42245 and the sixth calculation module 42246 calculate the target feature vector and the spectrum stability index according to the updated matrix;

[0058] Finally, the judgment output module 42248 judges whether the preset convergence condition is met according to the comprehensive convergence value. If so, the final spectrum stability index SF is output; if not, the iteration is continued to update the smoothing coefficient until the final result is obtained;

[0059] Through these hierarchical operations, the system can accurately extract productive noise-related components from complex mixed signals, conduct noise level assessments, and issue warning signals in a timely manner, thereby effectively protecting the health and safety of workers.

[0060] Embodiment 2

[0061] like Figure 2 As shown, the second embodiment of the present invention provides a multi-node edge sensing sampling method based on environmental monitoring, including the following operation steps:

[0062] Step S10: When the staff is in the current noisy place, the portable sensor worn by each staff receives the radio frequency signal sent from the central transmitting unit in real time;

[0063] It should be noted that in the above-mentioned embodiment of the present application, a central transmitting unit is set in a noisy place to send a stable radio frequency signal covering the entire current noisy place in real time. When the staff is in the noisy place, the portable sensor worn by the staff will be in the radio frequency signal coverage range of the central transmitter, and the portable sensor will receive the radio frequency signal anytime and anywhere within the current coverage range; the radio frequency signal is a stable signal source, providing a reference signal for subsequent processing and analysis.

[0064] Step S20: Mixing the radio frequency signal with the sound signal generated in the current noise location to obtain a mixed signal;

[0065] It should be noted that in the above-mentioned embodiment of the present application, the sound signal cannot be directly processed by computer operation, but the portable sensor in the embodiment of the present application will mix the radio frequency signal and the sound signal through a nonlinear element to generate a new frequency component, namely the above-mentioned mixed signal; the spectrum information of the noise signal can be converted into a frequency range that is easier to analyze, thereby facilitating subsequent processing and analysis operations; that is, the sound signal and the radio frequency signal can be mixed into an electrical signal and sent out through the mixed signal, which is convenient for subsequent direct calculation and processing of the sound signal;

[0066] It should be noted that the sound signal in the above-mentioned embodiment of the present application includes noise signals generated by mechanical impact, friction, rotation, etc. in the current noise place, which have a significant impact on human health, and the sound signal also includes mechanical noise different from the noise signal generated by the staff during work and walking;

[0067] Step S30: performing Fourier transform processing and screening processing on the mixed signal to obtain a noise signal;

[0068] Step S40: Analyze the noise signal to obtain a noise level; store and display the noise level, and issue a warning signal to the current staff based on the noise level.

[0069] It should be noted that the above-mentioned embodiment of the present application monitors the noise situation in the production noise place in real time through the cooperation of the portable sensor worn by the staff and the central transmitting unit, and analyzes the noise signal through the mixing technology, and finally provides the noise level information and issues a warning;

[0070] During the specific operation, the radio frequency signal emitted by the above-mentioned central transmitting unit acts as a stable signal source, providing a stable and consistent signal source for the subsequent mixing process, ensuring that the system can accurately extract and analyze noise signals; by mixing the noise signal collected by the portable sensor anytime and anywhere with the reference radio frequency signal, the system can extract and process noise signals of different frequencies, providing a data basis for subsequent noise analysis; further, components related to productive noise are extracted from the mixed signal, ensuring that the system can accurately identify noise signals that have an impact on human health from complex mixed signals, laying the foundation for subsequent noise level assessment; the noise level is analyzed according to the intensity or spectral characteristics of the noise signal. When the noise level exceeds the set threshold, the system promptly issues a warning signal to remind staff to pay attention to noise hazards and avoid long-term exposure to harmful noise environments; by providing real-time noise level assessment and warnings, the system can not only monitor the noise environment, but also issue safety warnings to staff to ensure that they work within a healthy range.

[0071] Specifically, in step S30, the mixed signal is subjected to Fourier transform processing and screening processing to obtain a noise signal, which includes the following operation steps:

[0072] Step S31: Perform Fourier transform processing based on the mixed signal to obtain spectrum information S y (f);

[0073] It should be noted that the above embodiment of the present application processes the mixed signal through Fourier transform to obtain the spectrum information S y (f), the frequency components of the noise can be extracted, and the amplitude distribution in different frequency regions can be analyzed; due to the different spectral characteristics of mechanical noise and quasi-mechanical noise (quasi-mechanical noise is the noise generated by the friction of clothing or appliances, conversation, etc. of workers during work and walking, as well as near the portable sensor), usually, mechanical noise is manifested as a relatively stable frequency component in the spectrum, while quasi-mechanical noise usually has a strong instantaneous variability in the spectrum, and the frequency distribution is relatively messy and periodic; therefore, the technical solution adopted in the embodiment of the present invention can more easily distinguish mechanical noise from quasi-mechanical noise through Fourier transform processing; the study found that it is very important to distinguish mechanical noise from quasi-mechanical noise when analyzing the noise field, which lays a technical foundation for the subsequent accurate screening and identification of noise types (especially identification of mechanical noise);

[0074] Step S32: Perform a time domain feature extraction operation on the mixing signal of the observation period (i.e., continuous time period) to obtain time domain feature information; based on the time domain feature information, filter the mixing signal to obtain target spectrum information S y (f');

[0075] The time domain feature information includes the root mean square RMS value RMS(y(t)) and the autocorrelation function R y (τ);

[0076] The root mean square RMS value RMS(y(t)) is calculated as follows:

[0077]

[0078] In the formula, represents the frequency of the mixed signal; T is the observation period (that is, the above observation period (i.e., continuous time period), or the time length of the mixed signal measurement);

[0079] The autocorrelation function R y (τ) is calculated as:

[0080]

[0081] In the formula, represents the expected value, τ is the time delay (or time offset, which represents the time difference between two time points of the mixing signal), and y(t) is the mixing signal at time point t;

[0082] It should be noted that, in the above-mentioned embodiment of the present application, the power or energy of the signal is measured by calculating the root mean square RMS value RMS(y(t)), which represents the average level of the signal amplitude. Under normal circumstances, mechanical noise usually presents a relatively stable root mean square RMS value RMS(y(t)), while quasi-mechanical noise usually has a higher instantaneous volatility; therefore, when the root mean square RMS value RMS(y(t)) exceeds the preset root mean square RMS value threshold, it is proved that the signal of a certain section (that is, the above-mentioned time period) in the mixed signal is quasi-mechanical noise; and when the root mean square RMS value RMS(y(t)) does not exceed the preset root mean square RMS value threshold, it is proved that the signal of a certain section (that is, the above-mentioned time period) in the mixed signal is mechanical noise, that is, the above-mentioned target spectrum information S y (f');

[0083] Through the above autocorrelation function R y (τ) is calculated to describe the correlation of the mixing signal y(t) at different time delays τ and to determine the periodicity of the signal. Generally, mechanical noise usually has a relatively stable autocorrelation feature, while quasi-mechanical noise usually exhibits irregular or transient autocorrelation features. Therefore, in the above autocorrelation function R y When (τ) exceeds the preset autocorrelation function, it is proved that the signal of a certain section (that is, the above time period) of the mixing signal is a mechanical noise; y(τ) does not exceed the preset autocorrelation function, it is proved that the signal of a certain section (that is, the above time period) in the mixed signal is mechanical noise, that is, the above target spectrum information S y (f').

[0084] Step S33: Based on the target spectrum information S y (f') Acquire multiple spectrum ranges; Based on the preset mechanical noise spectrum S mech (f) and the mechanical noise spectrum S class (f) Combine the spectrum range to obtain the typical mechanical noise spectrum Score mech Compared with the typical mechanical noise spectrum Score class ;

[0085] The typical mechanical noise spectrum Score mech The calculation method is:

[0086]

[0087] The typical mechanical noise spectrum Score class The calculation method is:

[0088]

[0089] In the formula, f max With f min is the spectrum range; df is the frequency interval;

[0090] It should be noted that the above embodiment of the present application first obtains the target spectrum information S y The peak values ​​f of (f') max and the peak-to-valley value f min , and then the peak value f max and the peak-to-valley value f min Sort by time series, then by adjacent peak values ​​f max and the peak-to-valley value f min Construct the spectrum range;

[0091] Step S34: Based on the typical mechanical noise spectrum Score mech Compared with the typical mechanical noise spectrum Score class And the time domain characteristic information is screened and analyzed to obtain a noise signal;

[0092] Specifically, in the above embodiment of the present application, the typical mechanical noise spectrum Score is determined by mech Compared with the typical mechanical noise spectrum Score class Specifically, the typical mechanical noise spectrum Score is judged.mech Is the value greater than the typical mechanical noise spectrum Score? class If yes, then determine the target spectrum information S corresponding to the spectrum range y (f') is the mechanical noise signal to be determined; if not, the typical mechanical noise spectrum Score is determined mech Is the value less than the typical mechanical noise spectrum Score? class If yes, then determine the target spectrum information S corresponding to the spectrum range y (f') is the mechanical noise signal to be determined;

[0093] It should be noted that the above-mentioned embodiment of the present application is screened twice. First, the signal is converted from the time domain to the frequency domain through Fourier transform, which lays the foundation for subsequent noise analysis. Mechanical noise and quasi-mechanical noise have different manifestations in the frequency domain. Fourier transform can help the system quickly identify these differences. The frequency composition of the signal can be directly observed through frequency domain information, which simplifies the complexity of time domain analysis and can more intuitively identify the frequency characteristics of the noise. Furthermore, according to the root mean square RMS value reflecting the energy level of the signal, it can effectively distinguish mechanical noise (stable energy) and quasi-mechanical noise (large energy fluctuations). This feature helps to quickly screen the noise in time. At the same time, it is also necessary to provide a deep understanding of the periodic characteristics of the signal through the autocorrelation function. Mechanical noise usually has a stable periodicity, while quasi-mechanical noise is more irregular. Through autocorrelation analysis, the system can further confirm the periodic characteristics of the signal and enhance the accuracy of noise classification; further, in the process of executing step S33-step S34, through the analysis of the spectrum peaks and peak-to-valley values, combined with the typical noise spectrum information, the spectrum range of mechanical noise and quasi-mechanical noise is more finely delineated. Through peak sorting and spectrum range construction, the system can more accurately define the noise type, avoiding rough classification based solely on frequency range, and improving the accuracy and reliability of noise identification; finally, by comparing the size of the typical mechanical noise spectrum and the quasi-mechanical noise spectrum, the noise type is finally confirmed. Such multi-level screening not only takes into account the spectrum characteristics, but also integrates the time domain analysis results, making the noise classification more accurate; through precise feature matching, the classification error is minimized, making the final noise identification result more reliable.

[0094] During the specific implementation of the above embodiment of the present application, the technicians also found that during the mechanical production process, the start-up of the test machine will cause the mechanical short-term friction to generate instantaneous noise, and in the above method of obtaining the target spectrum information S through the time domain feature information y (f'), this instantaneous noise will be included in the category of mechanical noise. Therefore, the target spectrum information S is obtained here through the time domain feature information y(f'), further analysis and identification of mechanical noise is required to obtain more accurate target spectrum information S y (f').

[0095] Specifically, in step S32, the mixed signal is screened based on the time domain feature information to obtain target spectrum information S y (f'), comprising the following steps:

[0096] Step S321: Performing a preliminary screening on the mixed signal based on the time domain feature information to obtain the first target spectrum information S y (f')1 and the mixed signal to be screened;

[0097] It should be noted that the above embodiment of the present application is to make the root mean square RMS value RMS(y(t)) less than or equal to the preset root mean square RMS value threshold, and at the same time in the autocorrelation function R y The mixing signal of the corresponding time period when (τ) is less than or equal to the preset autocorrelation function is determined as the first target spectrum information S y (f')1; and remove the first target spectrum information S y The mixed signals other than (f')1 are determined as mixed signals to be screened, so that the screening operation can be performed again later; the above-mentioned initial screening is the first screening operation, also known as the rough screening operation, which aims to screen the mixed signals that can be more directly identified as the target spectrum information S y (f') the first target spectrum information S y (f')1, and screen to obtain the mixed frequency signal to be screened, providing a data basis for the subsequent second screening (or fine screening) operation.

[0098] Step S322: re-screening the mixed signal to be screened based on the time dimension and the frequency component of the mixed signal to be screened to obtain the second target spectrum information S y (f')2;

[0099] It should be noted that, under normal circumstances, quasi-mechanical noise (quasi-mechanical noise is noise generated by friction between clothing or appliances of workers at work and walking, as well as noise generated by conversation, etc. near portable sensors) will appear as a continuous, low-frequency (low in the audio band and vibration band, and its unit is Hz) mixed-frequency signal, and starting the test machine will cause the transient noise generated by the short-term mechanical friction to be a high-frequency mixed-frequency signal; in addition, the quasi-mechanical noise is often a continuous mixed-frequency signal with a higher frequency than the transient noise generated by the short-term mechanical friction in terms of the frequency of occurrence; based on this, the embodiment of the present application screens the time dimension to make the transient noise generated by the short-term mechanical friction (that is, the second target spectrum information S y (f')2) can be screened out more accurately.

[0100] Step S323: Based on the first target spectrum information S y (f')1 and the second target spectrum information S y (f')2 Get the target spectrum information S y (f').

[0101] Explanation: The above-mentioned embodiment of the present application first uses time domain characteristics (i.e., root mean square RMS value and autocorrelation function) to perform a preliminary screening of the mixed signal, determines to obtain part of the mixed signal that can be directly identified as the target spectrum information (i.e., the first target spectrum information), and determines the remaining mixed signal as the signal to be screened, and quickly filters out the signal that meets the target characteristics through simple time domain characteristic analysis, which provides a basis for subsequent screening;

[0102] However, since the first target spectrum information can be directly screened, the second target spectrum information S y (f')2 is very important reference information, which contains some short-term mechanical noise that has not been identified into the target spectrum information, and it needs to be further shared and screened to be obtained. Therefore, in the above-mentioned embodiment of the present application, the mixed signal to be screened is further screened again, especially through the analysis of the time dimension and frequency component, so as to more accurately distinguish between mechanical noise and quasi-mechanical noise, thereby extracting the second target spectrum information, and accurately identifying high-frequency and short-term mechanical noise through refined time dimension and frequency component analysis, further improving the screening accuracy; finally, the first target spectrum information and the second target spectrum information are combined to obtain the final target spectrum information, and finally the results of the two screenings are combined to form accurate target spectrum information, thereby ensuring the comprehensiveness and accuracy of the final result.

[0103] Specifically, in step S322, a re-screening operation is performed based on the time dimension of the mixed signal to be screened and the frequency component of the mixed signal to be screened to obtain the second target spectrum information S y (f')2, comprising the following steps:

[0104] Step S3221: Divide the mixed frequency signal to be screened according to the time sequence corresponding to the mixed frequency signal to be screened to obtain a plurality of time windows i={i1, i2, i3, ..., iN}; wherein iN is the Nth time window;

[0105] Step S3222: performing short-time Fourier transform processing on each of the time windows i to obtain frequency components j={j1, j2, j3, ...jK} corresponding to each of the time windows i; wherein jK is the frequency component of the Kth time window;

[0106] Step S3223: constructing a time-frequency matrix M based on the time window i and the frequency components corresponding to the time window i;

[0107]

[0108] In the formula, m iN,jK represents the amplitude of the kth frequency component j corresponding to the Nth time window i;

[0109] It should be noted that the above-mentioned embodiments of the present application can simultaneously capture the time domain and frequency domain characteristics of the signal in the form of a matrix, which is convenient for analyzing the distribution characteristics of noise in the time-frequency plane.

[0110] Step S3224: Calculate the time-related index TC(i) based on the time-frequency matrix M;

[0111] The time-related index TC(i) is calculated as follows:

[0112] TC(i)=corr(M[i,:],M[i+1,:]);

[0113] In the formula, corr() is the correlation coefficient calculation function; M[i,:] is the i-th column element in the time-frequency matrix;

[0114] It should be noted that the above-mentioned embodiment of the present application measures the correlation of the spectrum of the mixed signal to be screened in the neighboring time window by calculating the time-related index; the spectrum characteristics of mechanical noise should have a strong correlation in adjacent time windows. By calculating the above-mentioned time-related index, mechanical noise-like noise with sudden changes in spectrum characteristics can be identified.

[0115] Step S3225: performing iterative calculation based on the time-frequency matrix M and in combination with the time scale and frequency scale of the time-frequency matrix M to obtain a spectrum stability index SF; calculating a comprehensive scoring function Score(i) based on the spectrum stability index SF and the time-related index TC(i);

[0116] The calculation method of the comprehensive scoring function Score(i) is:

[0117] Score(i)=w1 * TC(i)+w2 * SF(i);

[0118] In the formula, w1 and w2 are weight coefficients;

[0119] It should be noted that the spectrum distribution of mechanical noise should be relatively concentrated. By calculating the above-mentioned spectrum stability index SF, mechanical-like noise with discrete spectrum distribution can be distinguished.

[0120] Step S3226: Determine whether the comprehensive scoring function Score(i) is greater than or equal to a preset comprehensive scoring function threshold. If so (if not, determine that the mixed signal to be screened corresponding to the time window is mechanical noise and screen it out and no longer proceed to the next step), then determine that the mixed signal to be screened corresponding to the time window is the second target spectrum information S y (f')2.

[0121] It should be noted that, through the technical solution adopted in the above-mentioned embodiment of the present application, continuous mechanical noise and transient noise can be effectively distinguished, the accuracy of target spectrum information can be improved, and the misjudgment rate can be reduced, especially the misjudgment of friction noise generated at the moment of startup;

[0122] The above-mentioned embodiment of the present application first divides the mixed signal into multiple time windows and performs local analysis on the mixed signal to avoid the adverse effects of time variability that may exist when processing the entire signal. The signal features in each time window can be processed independently, which is convenient for capturing local noise patterns and spectral features. According to the above-mentioned embodiment of the present application, the changes of the signal in each time period can be effectively captured, especially for non-stationary signals (such as mechanical noise), the divided signal can reflect its temporal changes; further, the signal of each time window is converted from the time domain to the frequency domain through short-time Fourier transform, so that the spectral features in each time period can be clearly displayed. Further, by constructing a time-frequency matrix, the spectral information of each time window can be organized in a matrix form, which is convenient for a comprehensive analysis of the time-frequency characteristics of the signal; by calculating the time-related index, the quasi-mechanical noise with sudden changes in spectral features can be effectively distinguished, and the target spectral information can be further screened out; by calculating the spectral stability and time-related indexes, the mechanical noise with concentrated and stable spectral distribution and the quasi-mechanical noise with discrete spectral distribution and rapid changes can be accurately distinguished; finally, by threshold judgment, the signal that meets the mechanical noise characteristics can be further effectively screened out.

[0123] Specifically, in step S3225, an iterative calculation is performed based on the time-frequency matrix M and in combination with the time scale and frequency scale of the time-frequency matrix M to obtain a spectrum stability index SF, including the following operation steps:

[0124] Step S33251: constructing an initial spectrum stability matrix SSM(0) based on the time-frequency matrix M; calculating a density index D(i) based on the time-frequency matrix M; and initializing iteration parameters;

[0125] The iteration parameters include a smoothing coefficient α, an initial spectrum stability index SF(i), a time scale parameter τt, a frequency scale parameter τf, a weight coefficient β1, a weight coefficient β2, a weight coefficient β3, a learning rate γ (controlling the speed of adaptive adjustment of parameters), an iteration counter, and a maximum iteration threshold; the iteration number of the iteration counter is initially 0;

[0126] The initial spectrum stability matrix SSM(0) is expressed as:

[0127]

[0128] In the formula, m[i N ,j K ] represents the amplitude of the Kth frequency component j in the Nth time window i; μ K represents the mean value of the Kth frequency component (representing the average value of the frequency component in all time windows); σ K represents the standard deviation of the Kth frequency component (reflecting the degree of fluctuation of the Kth frequency component); ε is the smoothing factor (to prevent the denominator from being 0);

[0129] The density index D(i) is calculated as follows:

[0130] D(i)=sum(exp(-||M[i,:]-M[j,:]|| / σd));

[0131] Where, M[i,:] represents the i-th column element in the time-frequency matrix M (i.e., all frequency components in the same time window i); M[j,:] represents the j-th column element in the time-frequency matrix M (i.e., different time windows corresponding to the same frequency component j); σd is the density calculation bandwidth parameter (controlling the smoothness of the density estimation);

[0132] Step S33252: constructing a time domain similarity matrix U based on the time window and the time scale parameter τt; and constructing a frequency domain similarity matrix V based on the frequency component and the frequency scale parameter τf;

[0133] The time domain similarity matrix U is expressed as:

[0134]

[0135] The frequency domain similarity matrix V is expressed as:

[0136]

[0137] It should be noted that, in the above-mentioned embodiment of the present application, the time domain similarity matrix U represents the degree of correlation between different time windows, |i x -i NThe smaller | is, the higher the similarity in the time domain is; the frequency domain similarity matrix V represents the degree of correlation between different frequency components, |j x -j N The smaller | is, the higher the frequency domain similarity is;

[0138] Step S33253: Iteratively update the spectrum stability matrix: add 1 to the iteration number of the iteration counter to obtain the current iteration number; and calculate based on the smoothing coefficient α, the initial spectrum stability matrix SSM(0), the time domain similarity matrix U and the frequency domain similarity matrix V to obtain the updated spectrum stability matrix SSM(z+1);

[0139] The updated spectrum stability matrix SSM(z+1) is expressed as:

[0140] SSM(z+1)=α*SSM(z)+(1-α)*U*SSM(z)*V;

[0141] Wherein, z is the number of iterations of the previous iteration; (z+1) is the number of current iterations; SSM(z+1) represents the updated spectrum stability matrix obtained by updating the spectrum stability matrix generated by the previous iteration number under the current iteration number; SSM(z) represents the updated spectrum stability matrix under the previous iteration number (at the first iteration, that is, when the current iteration number is 1, SSM(z) is SSM(0), which represents the initial spectrum stability matrix);

[0142] Step S33254: Calculate and obtain the target eigenvector C(i) by solving the eigenvalue equation based on the updated spectrum stability matrix SSM(z+1);

[0143] It should be noted that the above embodiment of the present application solves the eigenvalue equation of the updated spectrum stability matrix SSM(z+1), and then takes the eigenvector corresponding to the maximum eigenvalue as the target eigenvector C(i);

[0144] Step S33255: Calculate the updated spectrum stability index SF based on the target feature vector C(i), the density index D(i) and the initial spectrum stability index SF(i) new (i);

[0145] The updated spectrum stability indicator SF new (i) is calculated as follows:

[0146] SF new (i)=β1*SF(i)+β2*C(i)+β3*D(i);

[0147] Step S33256: Calculate the target convergence threshold at the current number of iterations based on the current number of iterations combined with the preset initial convergence threshold; and calculate the target convergence threshold at the current number of iterations based on the updated spectrum stability index SF new (i) and the updated spectrum stability matrix SSM(z+1) to calculate the comprehensive convergence value TC;

[0148] Step S33257: Determine whether the comprehensive convergence value is less than or equal to the target convergence threshold. If so, output the updated spectrum stability index SF new (i) is the target spectrum stability index SF; if not, determining whether the current number of iterations is greater than or equal to a preset maximum number of iterations threshold, and if so, outputting the updated spectrum stability index SF new (i) is the target spectrum stability index SF; if not, the smoothing coefficient α is updated based on the updated spectrum stability matrix SSM(z+1) to obtain the updated smoothing coefficient α'; the updated smoothing coefficient α' is used as the smoothing coefficient α, and the iterative update spectrum stability matrix operation of the above step S33253 is returned and re-executed until the target spectrum stability index SF is output.

[0149] It should be noted that the above-mentioned embodiment of the present application first constructs an initial spectrum stability matrix SSM(0) based on the time-frequency matrix M, calculates the density index, and initializes the relevant iteration parameters, which lays the foundation for subsequent iterative updates and ensures that the spectrum stability matrix can be effectively updated during the iteration process; further, the distribution characteristics of the signal in the time and frequency dimensions are analyzed by the time and frequency domain similarity matrix, so as to provide support for subsequent spectrum stability updates and feature extraction; iterative optimization is performed through the information of the smoothing coefficient, the time domain and the frequency domain similarity matrix, so that the spectrum stability matrix gradually reflects the true stability characteristics of the signal, and the time and frequency similarity matrix is ​​used to optimize the spectrum matrix in combination with the smoothing coefficient to ensure that the matrix can effectively converge in both the time and frequency dimensions, and the target eigenvalue equation is used to extract the target eigenvector, and the spectrum stability characteristics of the signal are extracted as a concise and effective vector representation, which provides a basis for subsequent stability index updates; by calculating the updated spectrum stability index, the stability of the signal is further quantified, so as to finally screen out the signal characteristics that meet the target spectrum stability standard; further, by dynamically calculating the convergence value and the target convergence threshold, it is ensured that the iterative process can effectively converge, thereby avoiding unnecessary waste of iterative computing power, and finally outputting the target spectrum stability index SF.

[0150] Specifically, in step S33256, based on the updated spectrum stability indicator SF new(i) and the updated spectrum stability matrix SSM(z+1) to calculate the comprehensive convergence value TC, including the following steps:

[0151] Step S332561: Calculate the relative change rate RC of the spectrum stability matrix based on the updated spectrum stability matrix SSM(z+1); Calculate the relative change rate RC of the spectrum stability matrix based on the updated spectrum stability index SF new (i) Calculating the relative change rate SC of the stability index;

[0152] The calculation method of the relative change rate RC of the spectrum stability matrix is:

[0153]

[0154] Where SSM(z) is the updated spectrum stability matrix under the last iteration number;

[0155] The calculation method of the relative change rate SC of the stability index is:

[0156]

[0157] Where SF new (i)' is the updated spectrum stability index under the last iteration number (in the first iteration, SF new (i)' is the initial spectrum stability index SF(0));

[0158] Step S332562: Calculate a comprehensive convergence value TC based on the relative change rate RC of the spectrum stability matrix and the relative change rate SC of the stability index;

[0159] The calculation method of the comprehensive convergence value TC is:

[0160] TC = w3 × RC + w4 × SC;

[0161] In the formula, w3 and w4 are weight coefficients, and w1+w2=1;

[0162] It should be particularly noted that the above-mentioned embodiment of the present application effectively measures the degree of change of the spectrum stability matrix and the stability index of the system during the iteration process by calculating the relative change rate RC of the spectrum stability matrix and the relative change rate SC of the spectrum stability index, and determines the termination or continuation of the iteration through the comprehensive convergence value TC, thereby ensuring that the calculation can be stopped in time when the set convergence standard is reached to avoid unnecessary waste of calculation; RC and SC provide different perspectives on the local and global stability of the signal, and the comprehensive convergence value TC combines these two indicators in a weighted manner to provide a comprehensive convergence judgment. This method can be flexibly adjusted according to different signal characteristics, ensuring that the convergence progress of different stages in the iteration process can be effectively controlled.

[0163] Specifically, in step S33257, the smoothing coefficient α is updated based on the updated spectrum stability matrix SSM(z+1) to obtain an updated smoothing coefficient α', including the following operation steps:

[0164] Step S332571: Calculate the updated smoothing coefficient α' based on the relative change rate RC of the spectrum stability matrix;

[0165] The updated smoothing coefficient α' is calculated as follows:

[0166] α'=α×(1+γ×RC);

[0167] In the formula, γ is the learning rate;

[0168] It should be noted that the above-mentioned embodiment of the present application takes into account the relative change rate RC of the spectral stability matrix SSM(z+1) after the update, and introduces a dynamic adjustment mechanism for the smoothing coefficient by combining RC and the learning rate γ. The degree of change of the spectral stability matrix directly affects the update of the smoothing coefficient, which enables the smoothing process to be automatically adjusted according to the stability of the signal.

[0169] In summary, the multi-node edge sensing sampling system and method based on environmental monitoring proposed in the example of the present invention first uses the radio frequency signal transmitted by the central transmitting unit as a stable signal source, which provides a stable and consistent signal source for the subsequent mixing process, ensuring that the system can accurately extract and analyze noise signals; by mixing the noise signal collected by the portable sensor anytime and anywhere with the reference radio frequency signal, the system can extract and process noise signals of different frequencies, providing a data basis for subsequent noise analysis;

[0170] Furthermore, the components related to productive noise are extracted from the mixed signal, ensuring that the system can accurately identify noise signals that have an impact on human health from complex mixed signals, laying the foundation for the subsequent noise level assessment. The noise level is analyzed based on the intensity or spectrum characteristics of the noise signal. When the noise level exceeds the set threshold, the system promptly issues a warning signal to remind workers to pay attention to noise hazards and avoid long-term exposure to harmful noise environments. By providing real-time noise level assessment and warnings, the system can not only monitor the noise environment, but also issue safety warnings to workers to ensure that they work within a healthy range.

[0171] Furthermore, when extracting components related to productive noise from the mixed signal, through multi-level screening, both the spectrum characteristics and the time domain analysis results are considered, making the noise classification more accurate; through precise feature matching, the classification error is minimized, making the final noise identification result more reliable;

[0172] Furthermore, the time dimension and frequency components are analyzed and screened to form accurate target spectrum information, ensuring the comprehensiveness and accuracy of the final result;

[0173] Furthermore, by constructing a time-frequency matrix and calculating time-related indicators and spectrum stability, we can accurately distinguish between mechanical noise with a concentrated and stable spectrum distribution and mechanical-like noise with a discrete and rapidly changing spectrum distribution; finally, through threshold judgment, we can further effectively screen out signals that meet the characteristics of mechanical noise;

[0174] Furthermore, in the specific operation process, the initial spectrum stability matrix SSM(0) constructed based on the time-frequency matrix M, the density index and the iteration parameters are calculated, and the comprehensive convergence value and the smoothing coefficient that change dynamically during the iteration process are combined to dynamically judge and output the target spectrum stability index SF.

[0175] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. A person skilled in the art may modify the technical solutions described in the above embodiments, or replace part or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-node edge sensing sampling system based on environmental monitoring, characterized in that: It includes a central transmitting unit, a data receiving module, a mixing module, a processing and screening module, and an analysis and warning module; The central transmitting unit is used to transmit radio frequency signals in real time in the current noisy place; The data receiving module is used to receive the radio frequency signal sent from the central transmitting unit in real time in the current noise place; the data receiving module is a sensor with a signal receiver function worn by each individual; The mixing signal is used to mix the radio frequency signal with the sound signal generated by the current noise place to obtain a mixing signal; The processing and screening module is used to perform Fourier transform processing and screening processing on the mixed signal to obtain a noise signal; The analysis and warning module is used to analyze the noise signal to obtain the noise level; The noise level is stored and displayed, and a warning signal is issued to the current staff based on the noise level.

2. A multi-node edge sensing sampling system based on environmental monitoring according to claim 1, characterized in that: The processing and screening module includes a transformation processing module, a feature extraction and screening module, a matching module and a screening analysis module; The transform processing module is used to perform Fourier transform processing based on the mixed signal to obtain spectrum information S y (f); The feature extraction and screening module is used to perform a time domain feature extraction operation on the mixing signal of the observation period T to obtain time domain feature information; based on the time domain feature information, the mixing signal is screened to obtain target spectrum information S y (f'); The matching module is used to y (f') Acquire multiple spectrum ranges; Based on the preset mechanical noise spectrum S mech (f) and the mechanical noise spectrum S class (f) Combine the spectrum range to obtain the typical mechanical noise spectrum Score mech Compared with the typical mechanical noise spectrum Score class ; The screening and analysis module is used to mech Compared with the typical mechanical noise spectrum Score class And the time domain characteristic information is screened and analyzed to obtain a noise signal.

3. A multi-node edge sensing sampling system based on environmental monitoring according to claim 2, characterized in that: The time domain feature information includes the root mean square RMS value RMS(y(t)) and the autocorrelation function R y (τ).

4. A multi-node edge sensing sampling system based on environmental monitoring according to claim 3, characterized in that: The screening and analysis module is specifically used to determine the typical mechanical noise spectrum Score mech Is the value greater than the typical mechanical noise spectrum Score? class If yes, then determine the target spectrum information S corresponding to the spectrum range y (f') is the mechanical noise signal to be determined; if not, the typical mechanical noise spectrum Score is determined mech Is the value less than the typical mechanical noise spectrum Score? class If yes, then determine the target spectrum information S corresponding to the spectrum range y (f') is the mechanical noise signal to be determined.

5. A multi-node edge sensing sampling system based on environmental monitoring according to claim 4, characterized in that: The feature extraction and screening module includes a primary screening module, a re-screening module and a merging module; The primary screening module is used to perform a primary screening on the mixed signal based on the time domain feature information to obtain the first target spectrum information S y (f')1 and the mixed signal to be screened; The re-screening module is used to perform a re-screening operation based on the time dimension of the mixed signal to be screened and the frequency component of the mixed signal to be screened to obtain the second target spectrum information S y (f')2; The merging module is used to y ( f' ) 1 and the second target spectrum information S y ( f' ) 2 Obtain target spectrum information S y ( f' ) .

6. A multi-node edge sensing sampling system based on environmental monitoring according to claim 5, characterized in that: The re-screening module includes a time window division module, a frequency component acquisition module, a first calculation module, a second calculation module and a judgment module; The time window division module is used to divide the mixed frequency signal to be screened according to the time sequence corresponding to the mixed frequency signal to be screened to obtain a plurality of time windows i={i1, i2, i3, ...iN}; wherein iN is the Nth time window; The frequency component acquisition module is used to obtain the frequency components j={j1, j2, j3, ...jK} corresponding to each time window i by short-time Fourier transform processing; wherein jK is the frequency component of the Kth time window; The first calculation module is used to calculate the time-related index TC based on the time-frequency matrix M. ( i ) ; The second calculation module is used to iteratively calculate the spectrum stability index SF based on the time-frequency matrix M and in combination with the time scale and frequency scale of the time-frequency matrix M; based on the spectrum stability index SF and the time-related index TC ( i ) Calculate the comprehensive scoring function Score ( i ) ; The judgment module is used to judge the comprehensive scoring function Score ( i ) Is it greater than or equal to a preset comprehensive scoring function threshold? If so, it is determined that the mixed signal to be screened corresponding to the time window is the second target spectrum information S y ( f' ) 2.

7. A multi-node edge sensing sampling system based on environmental monitoring according to claim 6, characterized in that: The second calculation module includes an initialization module, a matrix construction module, a third calculation module, a fourth calculation module, a fifth calculation module, a sixth calculation module, a seventh calculation module and a judgment output module; The initialization module is used to construct an initial spectrum stability matrix SSM(0) based on the time-frequency matrix M; calculate a density index D(i) based on the time-frequency matrix M; and initialize iteration parameters; The matrix construction module is used for the iteration parameters including a smoothing coefficient α, an initial spectrum stability index SF(i), a time scale parameter τt, a frequency scale parameter τf, a weight coefficient β1, a weight coefficient β2, a weight coefficient β3, a learning rate γ, an iteration counter, and a maximum iteration number threshold; the iteration number of the iteration counter is initially 0; The third calculation module is used to construct a time domain similarity matrix U based on the time window and the time scale parameter τt; and to construct a frequency domain similarity matrix V based on the frequency component and the frequency scale parameter τf; The fourth calculation module is used for iteratively updating the spectrum stability matrix: adding 1 to the iteration number of the iteration counter to obtain the current iteration number; and calculating based on the smoothing coefficient α, the initial spectrum stability matrix SSM(0), the time domain similarity matrix U and the frequency domain similarity matrix V to obtain the updated spectrum stability matrix SSM(z+1); The fifth calculation module is used to calculate the target eigenvector C(i) by solving the eigenvalue equation based on the updated spectrum stability matrix SSM(z+1); The sixth calculation module is used to calculate the updated spectrum stability index SF based on the target feature vector C(i), the density index D(i) and the initial spectrum stability index SF(i). new (i); The seventh calculation module is used to calculate the target convergence threshold under the current number of iterations based on the current number of iterations combined with the preset initial convergence threshold; and based on the updated spectrum stability index SF new (i) and the updated spectrum stability matrix SSM(z+1) to calculate the comprehensive convergence value TC; The judgment output module is used to judge whether the comprehensive convergence value is less than or equal to the target convergence threshold. If so, the updated spectrum stability index SF is output. new (i) is the target spectrum stability index SF; if not, determining whether the current number of iterations is greater than or equal to a preset maximum number of iterations threshold, and if so, outputting the updated spectrum stability index SF new( i ) is the target spectrum stability index SF; if not, then based on the updated spectrum stability matrix SSM ( z+1 ) The smoothing coefficient α is updated to obtain an updated smoothing coefficient α'; the updated smoothing coefficient α' is used as the smoothing coefficient α, and the above-mentioned iterative update spectrum stability matrix operation is returned and re-executed until the target spectrum stability index SF is output.

8. A multi-node edge sensing sampling system based on environmental monitoring according to claim 7, characterized in that: The seventh calculation module is specifically configured to calculate the spectrum stability matrix SSM based on the updated spectrum stability matrix SSM. ( z+1 ) Calculate the relative change rate RC of the spectrum stability matrix; based on the updated spectrum stability index SF new( i ) Calculate the relative change rate SC of the stability index; The comprehensive convergence value TC is calculated based on the relative change rate RC of the spectrum stability matrix and the relative change rate SC of the stability index.

9. A multi-node edge sensing sampling system based on environmental monitoring according to claim 8, characterized in that: The judgment output module is specifically used to calculate the updated smoothing coefficient α' based on the relative change rate RC of the spectrum stability matrix; The updated smoothing coefficient α' is calculated as follows: α'=α×(1+γ×RC); Where γ is the learning rate.

10. A multi-node edge sensing sampling method based on environmental monitoring, characterized in that: The steps are as follows: When the staff is in the current noisy place, the portable sensor worn by each staff member receives the radio frequency signal sent from the central transmitting unit in real time; Based on mixing the radio frequency signal with the sound signal generated in the current noise place, a mixed signal is obtained; Performing Fourier transform processing and screening processing on the mixed signal to obtain a noise signal; Analyze the noise signal to obtain a noise level; The noise level is stored and displayed, and a warning signal is issued to the current staff based on the noise level.

Images