A method for processing electrical signals of a disease prevention environmental sensor

Abstract

The present invention proposes a method for processing electrical signals of a disease prevention environment sensor, which relates to the technical field of signal processing. The specific steps are as follows: making a dataset of electrical signals of the disease prevention environment, introducing a gain factor and combining graph gain and spatial distance to perform fusion processing on the electrical signals of the disease prevention environment, using multi-scale fractional-order differential and wavelet basis function decomposition to extract abnormal signal patterns, further enhancing abnormal components through manifold geometric gradient and fractional-order differential in reverse, introducing an adaptive threshold to calculate the abnormal confidence level, and fusing the confidence levels of different sensors through Riemannian manifold distance, combining a feedback mechanism to dynamically adjust parameters, and optimizing the processing of electrical signals of the disease prevention environment; at the same time, a model for processing electrical signals of the disease prevention environment is proposed, and the electrical signals of the disease prevention environment are optimized through the model; the model efficiently completes the detection and optimization of the electrical signals of the disease prevention environment, providing accurate data support for disease prevention.

Description

Technical Field

[0001] The present invention belongs to the technical field of signal processing, and particularly relates to a method for processing electrical signals of a disease prevention environment sensor. Background Art

[0002] In the field of disease prevention, the electrical signals of disease prevention environment sensors collected by environmental sensors provide important data support for health monitoring and early disease diagnosis; the electrical signals of disease prevention environment sensors can provide real-time health data to help detect disease risks in a timely manner; however, there are some technical bottlenecks in existing signal processing methods. When dealing with the electrical signals of disease prevention environment sensors, traditional signal processing methods are difficult to accurately extract useful information and are easily affected by noise, interference, and non-linear signal changes, resulting in unstable and inaccurate signal processing results; traditional anomaly detection methods usually rely on fixed thresholds and simple statistical models, and perform poorly in a dynamically changing environment and cannot adapt to rapidly changing health states and environmental conditions, leading to false alarms and missed detections, affecting the accuracy of disease prevention.

[0003] In addition, existing sensor data fusion methods do not fully consider the spatio-temporal relationship between sensors, and the accuracy after data fusion of different sensors is not high, and the monitoring ability of disease prevention cannot be effectively improved; in disease prevention, the parameter values of temperature and humidity, air quality, and gas concentration have a significant impact on human health, and the precise processing of the electrical signals of disease prevention environment sensors is directly related to the effect of disease prevention; the impact of environmental changes on human health is significant, and precisely processing the electrical signals of disease prevention environment, and timely discovering potential risks have become urgent problems to be solved in the field of disease prevention. Summary of the Invention

[0004] The present invention provides a method for processing electrical signals of a disease prevention environment sensor, aiming to propose a processing model for electrical signals of a disease prevention environment, in which a gain factor and a graph gain method are introduced to optimize the fusion of electrical signals of a disease prevention environment and improve signal quality and consistency; a multi-scale processing technology of fractional-order differential and wavelet basis functions is adopted to accurately extract the abnormal patterns of electrical signals of a disease prevention environment and calculate the initial abnormal probability; a hybrid Gaussian-Riemannian manifold space and manifold geometric gradient are used to enhance the abnormal components and improve the recognition accuracy of abnormal patterns of electrical signals of a disease prevention environment. By calculating the abnormal confidence and combining spatial consistency to generate an adaptive threshold, the detection of abnormal electrical signals of a disease prevention environment is optimized; the confidence levels of different sensors are fused based on the Riemannian manifold distance, and the parameters are dynamically adjusted to achieve efficient response in a complex environment; the processing model for electrical signals of a disease prevention environment is used to complete the accurate processing and anomaly detection of electrical signals of a disease prevention environment.

[0005] To achieve the above object, the present invention provides the following technical solution: A method for processing electrical signals of a disease prevention environment sensor, the specific steps are as follows:

[0006] S1. Collect the electrical signals collected by the environmental sensor and create a disease prevention environment electrical signal dataset;

[0007] S2. According to the spatio-temporal relationship of the disease prevention environment electrical signals, introduce a gain factor, and combine the graph gain method with the spatial distance to perform fusion processing on the disease prevention environment electrical signals;

[0008] S3. Generate multi-scale components by decomposing the signal through multi-scale fractional-order differential and wavelet basis functions, extract abnormal signal patterns, and calculate the initial abnormal probability by combining the sensor spatial distance and the electrical signal difference;

[0009] S4. Map the disease prevention environment electrical signals to the mixed Gaussian-Riemannian manifold space, perform enhancement processing on the disease prevention environment electrical signals, and reverse enhance the abnormal components through the manifold geometric gradient and fractional-order differential;

[0010] S5. Generate an adaptive threshold based on the mean, standard deviation, and spatial consistency measure of the local disease prevention environment electrical signals, and generate an abnormal confidence level by combining the enhanced disease prevention environment electrical signals and the initial abnormal probability;

[0011] S6. Fuse the confidence levels of different sensors through the Riemannian manifold distance, verify the abnormal patterns, and feedback the fusion result to the initial step S2, dynamically adjust the fractional-order differential parameters, perform feedback optimization on the disease prevention environment electrical signals, and iteratively execute until convergence to obtain the finally processed disease prevention environment electrical signals and the final fusion confidence level;

[0012] S7. Construct a disease prevention environment electrical signal processing model, input the disease prevention environment electrical signal dataset, sequentially go through steps S2 to S6, and complete the optimization processing of the disease prevention environment electrical signals through iterative training until convergence.

[0013] Preferably, in step S1, for the production of the disease prevention environmental electrical signal dataset, environmental sensors suitable for disease prevention monitoring are selected. The environmental sensors include temperature and humidity sensors, air quality sensors, and gas concentration sensors, which are mainly used to monitor environmental parameters such as temperature and humidity, air quality, and gas concentration. A nursing home is selected as the collection site for the disease prevention environmental electrical signal dataset. The sensors are installed at different monitoring positions to ensure coverage of various environmental conditions in the target area. The environmental sensors are used to continuously collect real-time data, record the electrical signals under different time and space conditions, and all the collected disease prevention environmental electrical signals will be stored together with the corresponding sensor sampling values and spatial coordinate information, and data annotation will be carried out. The disease prevention environmental electrical signals and the annotation information are sorted into a unified format to complete the production of the disease prevention environmental electrical signal dataset.

[0014] Preferably, in step S2, the specific steps for fusing the disease prevention environmental electrical signals are as follows:

[0015] S21. Input the original electrical signals of the sensors and the spatial coordinates of the sensors , which represents the sampling value of the i-th sensor node at time t, where is the total number of sensor nodes,

[0016] ;

[0017] In the formula, is the gain factor, representing the signal gain between sensors i and j at time t, and are the signal values of sensors i and j at time t respectively, is the exponential function, and are the adjustment coefficients;

[0018] S22. Adopt the graph gain method for signal fusion, construct the graph structure between the disease prevention environmental electrical signals and optimize the signal fusion process. The mathematical model for the fusion of the disease prevention environmental electrical signals is:

[0019] ;

[0020] In the formula, is the fusion signal matrix optimized by the graph gain, is the eigenvector matrix. By performing principal component analysis on the disease prevention environmental electrical signals, the first three principal components are extracted, and the energy of the principal components is retained. is the graph gain matrix. The graph gain matrix is a diagonal matrix, and the diagonal elements are gain factors. is the relationship matrix of the disease prevention environmental electrical signals. The relationship matrix is obtained by calculating the covariance matrix of the disease prevention environmental electrical signals. is the transpose matrix of is the element-wise multiplication;

[0021] S23. By combining the gain factor, graph gain, and spatial distance, and processing the signal using an exponential function, the optimized disease prevention environmental electrical signal is obtained. The mathematical model is:

[0022] ;

[0023] In the formula, is the optimized disease prevention environmental electrical signal, represents the position of sensor i, is the adjustment factor for controlling the spatial distance attenuation term, is the average value of the sensor positions.

[0024] Preferably, in step S2, according to the spatio-temporal relationship of the disease prevention environmental electrical signals, a gain factor is introduced, and the graph gain method and spatial distance are combined to optimize the fused disease prevention environmental electrical signal; First, the design of the gain factor adjusts the relative gain between the disease prevention environmental electrical signals by considering the spatio-temporal correlation between sensors, thereby improving the complementarity and effectiveness between the disease prevention environmental electrical signals; The mathematical model of the gain factor uses an exponential function to dynamically adjust the gain of the signal according to the change of the disease prevention environmental electrical signals, enhancing the sensitivity and adaptability of the signal; Second, the graph gain method is used to construct the relationship graph between the disease prevention environmental electrical signals, and the signal fusion is realized by optimizing the graph gain matrix, further improving the accuracy and robustness of the disease prevention environmental electrical signal processing; Finally, by combining the spatial distance attenuation term, the weighted fusion of the disease prevention environmental electrical signals is optimized by adjusting the influence of the spatial distance on the signal, making the fused disease prevention environmental electrical signal more accurate, effectively reducing the interference of the spatial position on the signal, and enhancing the monitoring ability of the disease prevention system; The overall design significantly improves the flexibility and accuracy of the disease prevention environmental electrical signal processing through the combination of the gain factor, graph gain, and spatial distance control, providing a more efficient monitoring means for disease prevention.

[0025] Preferably, in step S3, the specific steps for calculating the initial abnormal probability are:

[0026] S31. Decompose the disease prevention environment electrical signal into a multi-scale tensor, and directly extract the potential abnormal patterns. The mathematical model is:

[0027] ;

[0028] In the formula, is the component of the disease prevention environment electrical signal after multi-scale decomposition, is the fractional differential operator, , controlling the decomposition of the disease prevention environment electrical signal at different scales. s is the scale parameter, correspond to the high-frequency, medium-frequency, and low-frequency components respectively, is the adaptive wavelet basis function, controlling the frequency band components through the scale parameter s;

[0029] The mathematical model of the fractional differential operator is:

[0030] ;

[0031] In the formula, is the gamma function, normalizing the result of the fractional differential, is to take the integer part of, a is the historical time, and t is the current time point;

[0032] The mathematical model of the adaptive wavelet basis function is:

[0033] ;

[0034] S32. Combine the multi-scale components with the spatial correlation to generate the initial abnormal probability. The mathematical model is:

[0035] ;

[0036] In the formula, is the initial abnormal probability, is the Euclidean space distance between sensor nodes i and j, is the smoothing constant, is the variance of the disease prevention environment electrical signal at the scale parameter s.

[0037] Preferably, in step S3, the disease prevention environment electrical signal is decomposed by multi-scale fractional differential and wavelet basis functions to extract the abnormal patterns of the disease prevention environment electrical signal, and the initial abnormal probability is calculated by combining the spatial distance between sensors and the difference in the disease prevention environment electrical signal. First, the fractional differential is used to perform multi-scale processing on the disease prevention environment electrical signal to capture the multi-level features of the signal at different scales, which has better sensitivity to the subtle changes in the disease prevention environment electrical signal, thereby improving the accuracy of abnormal detection. The fractional differential operator flexibly controls the response of the disease prevention environment electrical signal in different frequency bands by adjusting the order, optimizing the time-domain and frequency-domain characteristics of the signal. The wavelet basis function is used to further decompose the disease prevention environment electrical signal to extract the potential abnormal patterns in the signal. By adjusting the adaptive wavelet basis function, precise processing of signals in different frequency bands is achieved, and the key features in the time-varying signal are captured. Finally, by combining the spatial distance between sensors and the difference in the disease prevention environment electrical signal, the initial abnormal probability is calculated to identify the abnormalities in the disease prevention environment electrical signal at different spatial positions, providing a reliable basis for abnormal detection and signal enhancement. The overall design improves the accuracy of the disease prevention environment electrical signal analysis and the sensitivity of abnormal detection through the combination of multi-scale decomposition and spatial correlation.

[0038] Preferably, in step S4, the specific steps for enhancing the disease prevention environment electrical signal are as follows:

[0039] S41. Map the disease prevention environment electrical signal to a mixed Gaussian-Riemannian manifold to enhance the abnormal components. The mathematical model is:

[0040] ;

[0041] In the formula, is the enhanced disease prevention environment electrical signal, is the number of Gaussian components, is the uniform mixing coefficient, is the manifold center, is the covariance, is the adaptive fractional parameter, , is the information entropy, is the stability constant to prevent the denominator from being zero, ;

[0042] S42. Enhance the abnormal signal by the fractional gradient in the reverse direction. The mathematical model is:

[0043] ;

[0044] In the formula, is the manifold geometric gradient calculated by the Riemann logarithm mapping, , is the logarithmic mapping on the Riemannian manifold, which maps the disease prevention environmental electrical signal from the manifold space to the tangent space. To enhance the intensity, , is of a value equal to 0.5, is the sign function, is the assignment symbol, indicating that the signal is updated by gradient enhancement.

[0045] Preferably, in step S4, by mapping the disease prevention environmental electrical signal to the hybrid Gaussian-Riemannian manifold space and using the manifold geometric gradient and fractional differential method to reverse enhance the abnormal components, the enhancement processing of the disease prevention environmental electrical signal is realized; First, map the disease prevention environmental electrical signal to the hybrid Gaussian-Riemannian manifold space, so that the disease prevention environmental electrical signal can be more effectively processed in a space with geometric structure, which helps to retain the inherent characteristics and spatial structure of the disease prevention environmental electrical signal; Through the calculation of the manifold geometric gradient, combined with the geometric properties of the manifold, enhance the abnormal components of the disease prevention environmental electrical signal, thereby improving the recognition accuracy of abnormal patterns; Use the reverse enhancement method of fractional differential to optimize the abnormal components in the disease prevention environmental electrical signal and enhance the recognizability; The design of fractional differential enables fine adjustment of the disease prevention environmental electrical signal at different scales, providing higher flexibility for the detail processing of the abnormal patterns of the disease prevention environmental electrical signal; The enhancement strategy in step S4 effectively improves the quality of the disease prevention environmental electrical signal, providing a clearer basis of the disease prevention environmental electrical signal for abnormal detection and pattern recognition.

[0046] Preferably, in step S5, the specific steps for calculating the abnormal confidence are:

[0047] S51. Generate an adaptive threshold based on the signal statistical characteristics and spatial consistency. The mathematical model is:

[0048] ;

[0049] In the formula, is the adaptive threshold, is the local signal mean, , is the local standard deviation, , is half the width of the time window, is the total length of the time window, and the calculation window is 5 seconds, is the spatial consistency metric, is the adaptive threshold amplitude adjustment coefficient, is the spatial consistency adjustment coefficient;

[0050] The mathematical model of the spatial consistency metric is as follows:

[0051] ;

[0052] In the formula, is the cross-correlation coefficient, which calculates the correlation of the disease prevention environment electrical signals;

[0053] The mathematical model of the cross-correlation coefficient is as follows:

[0054] ;

[0055] In the formula, is the average value of, is the average value of, is the summation over all moments in time t;

[0056] S52. Generate the confidence level by combining the jointly enhanced disease prevention environment electrical signals and the adaptive threshold. The mathematical model is as follows:

[0057] ;

[0058] In the formula, is the anomaly confidence level, is the normalization constant, , and max is to take the maximum value.

[0059] Preferably, in step S5, by calculating the anomaly confidence level of the disease prevention environment electrical signals, combining the statistical characteristics and spatial consistency of the local disease prevention environment electrical signals, an adaptive threshold is generated and the anomaly detection is further optimized. First, an adaptive threshold is generated through the mean value, standard deviation, and spatial consistency metric of the disease prevention environment electrical signals. The adaptive threshold can be adaptively adjusted according to the local fluctuations and spatial distribution of the disease prevention environment electrical signals, and can flexibly respond to the changes in the disease prevention environment electrical signals. The design of the adaptive threshold takes into account the time-varying characteristics of the disease prevention environment electrical signals and can accurately distinguish normal signals from abnormal signals at different time periods and spatial positions. By combining the enhanced disease prevention environment electrical signals with the initial anomaly probability, the anomaly confidence level is generated through weighted fusion, which can effectively improve the determination accuracy of the anomaly pattern. By introducing the spatial consistency metric, the correlation processing between the disease prevention environment electrical signals is further improved, and abnormal situations can also be correctly identified in complex environments. The overall design improves the sensitivity and robustness of the disease prevention environment electrical signal processing by combining the statistical characteristics and spatial consistency of the local disease prevention environment electrical signals, providing a reliable basis for accurate disease prevention.

[0060] Preferably, in step S6, the specific steps of fusing and feedback optimizing different disease prevention environment electrical signals are as follows:

[0061] S61. Fuse the electro-signals of different disease prevention environments based on information geometric distance to verify the abnormal pattern. The mathematical model is as follows:

[0062] ;

[0063] In the formula, is the fused confidence after fusing the confidences of different sensors, is the Riemannian manifold distance, , is the fusion attenuation coefficient, , and Var is the variance;

[0064] S62. The fused confidence is fed back to S2 to adjust the fractional differential parameter . The mathematical model is as follows:

[0065] ;

[0066] In the formula, the iteration termination condition is that the variance of is less than the preset threshold.

[0067] Preferably, by combining the confidence of the electro-signals of the disease prevention environment and using the Riemannian manifold distance to fuse the electro-signals of the disease prevention environment of different sensors, the verification and adjustment of the abnormal pattern are optimized. First, the electro-signals of the disease prevention environment collected by sensors at different spatial positions may have different degrees of errors and interferences. By using the Riemannian manifold distance to fuse the confidences of each sensor, the data from different sensors can be weighted and combined according to the spatial relationship. The fusion method based on geometric distance can improve the recognition accuracy of abnormal patterns in multi-source signals and can effectively enhance the sensitivity to environmental changes in the environmental monitoring scenario of the disease prevention system. The fused confidence result is fed back to step S2 to dynamically adjust the fractional differential parameter, which can adjust the processing parameters according to the detected abnormal electro-signals of the disease prevention environment during the real-time monitoring process to adapt to the current environmental state, thereby improving the adaptability and accuracy in different situations. The iterative optimization process gradually improves the accuracy of abnormal signal detection, ensuring that the disease prevention system can operate stably under complex and variable environmental conditions. The overall design improves the fusion accuracy and robustness of the electro-signals of the disease prevention environment by combining the Riemannian manifold distance and dynamic parameter adjustment, providing a strong technical guarantee for disease prevention and environmental monitoring.

[0068] Compared with the prior art, the present invention has the following technical effects: The technical solution provided by the present invention proposes a disease prevention environmental electrical signal processing model, which combines disease prevention environmental electrical signal fusion and anomaly detection technologies, significantly improving the overall monitoring ability of disease prevention. Among them, the introduction of a gain factor and a graph gain method optimizes the fusion process of disease prevention environmental electrical signals, enhances the effective information in the signals, and improves the accuracy and reliability of anomaly detection; The multi-scale processing technology of fractional-order differential and wavelet basis functions is adopted to accurately extract the anomaly patterns of disease prevention environmental electrical signals and calculate the initial anomaly probability, enhancing the anomaly detection ability of disease prevention environmental electrical signals and timely discovering potential health risks; The use of a mixture of Gaussian-Riemannian manifold space and manifold geometric gradient enhances the anomaly components, effectively improving the recognition accuracy of anomaly patterns in disease prevention environmental electrical signals. By calculating the anomaly confidence and combining spatial consistency to generate an adaptive threshold, high-precision anomaly detection is ensured; The confidence levels of different sensors are fused based on the Riemannian manifold distance to optimize the anomaly disease prevention environmental electrical signals; Using the disease prevention environmental electrical signal processing model, the deficiencies of traditional disease prevention methods in disease prevention environmental electrical signal processing and anomaly detection are solved, significantly improving the accuracy of monitoring results, ensuring that disease prevention can be carried out in a timely and effective manner, providing more accurate data support for disease prevention, reducing the risk of disease occurrence and transmission, and making the identification and response of health risks more efficient, thereby improving the overall health management level. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 is a flowchart of the disease prevention environmental electrical signal processing method provided by the present invention.

[0070] Figure 2 is a structural diagram of fusing disease prevention environmental electrical signals provided by the present invention.

[0071] Figure 3 is a structural diagram of calculating the initial anomaly probability provided by the present invention.

[0072] Figure 4 is a structural diagram of enhancing disease prevention environmental electrical signals provided by the present invention.

[0073] Figure 5 is a structural diagram of calculating the anomaly confidence provided by the present invention.

[0074] Figure 6 is a structural diagram of fusing and feedback optimizing different disease prevention environmental electrical signals provided by the present invention.

[0075] Figure 7 is a comparison diagram of the disease prevention environmental electrical signal with enhanced anomaly components after being processed by the disease prevention environmental electrical signal processing model and the original disease prevention environmental electrical signal provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0076] The present invention aims to propose a method for processing electrical signals of a disease prevention environment sensor, and propose a disease prevention environment electrical signal processing model, in which a gain factor and a graph gain method are introduced to optimize the fusion of disease prevention environment electrical signals, improve signal quality and consistency; adopt a multi-scale processing technology of fractional-order differentiation and wavelet basis functions to accurately extract abnormal patterns of disease prevention environment electrical signals and calculate the initial abnormal probability; use a mixture of Gaussian-Riemannian manifold space and manifold geometric gradient to enhance abnormal components, improve the recognition accuracy of abnormal patterns of disease prevention environment electrical signals, generate an adaptive threshold by calculating the abnormal confidence and combining spatial consistency, and optimize the detection of abnormal disease prevention environment electrical signals; fuse the confidence degrees of different sensors based on the Riemannian manifold distance, dynamically adjust parameters, and achieve efficient response in complex environments; use the disease prevention environment electrical signal processing model to complete the accurate processing and abnormal detection of disease prevention environment electrical signals.

[0077] Please refer to Figure 1 As shown, a method for processing electrical signals of a disease prevention environment sensor in an embodiment of the present application is as follows.

[0078] S1. Collect the electrical signals collected by the environmental sensor and make a disease prevention environment electrical signal dataset.

[0079] Further, in step S1, for the production of the disease prevention environment electrical signal dataset, select environmental sensors suitable for disease prevention monitoring. The environmental sensors include temperature and humidity sensors, air quality sensors, and gas concentration sensors, which are mainly used to monitor environmental parameters such as temperature and humidity, air quality, and gas concentration. Select a nursing home as the collection place for the disease prevention environment electrical signal dataset, install the sensors at different monitoring positions to ensure coverage of various environmental conditions in the target area, use the environmental sensors to continuously collect real-time data, record the electrical signals under different time and space conditions, and store all the collected disease prevention environment electrical signals together with the corresponding sensor sampling values and spatial coordinate information, and perform data annotation. Organize the disease prevention environment electrical signals and annotation information into a unified format to complete the production of the disease prevention environment electrical signal dataset.

[0080] S2. According to the spatio-temporal relationship of the disease prevention environment electrical signals, introduce a gain factor, and combine the graph gain method with the spatial distance to perform fusion processing on the disease prevention environment electrical signals.

[0081] Further, in step S2, for optimizing the fusion of environmental disease prevention environment electrical signals, its structure is as Figure 2 shown. The specific steps for performing fusion processing on the disease prevention environment electrical signals are as follows.

[0082] S21. Input the original disease prevention environment electrical signal and the spatial coordinates of the sensors , represents the sampling value of the i-th sensor node at time t, is the total number of sensor nodes. Set N to 50. 50 sensors continuously collect environmental data at different positions. The data collected by each sensor is a 1-hour time series signal with a sampling interval of 10 seconds. The total number of sampling points of the environmental electrical signal for disease prevention is 18,000, is the three-dimensional spatial coordinate of sensor i. After the original environmental electrical signal for disease prevention is processed by mean zeroing and variance normalization, the spatial and temporal relationships between the environmental electrical signals for disease prevention are calculated, and a correlation model of the signals is established to obtain the gain factor between the environmental electrical signals for disease prevention. The mathematical model of the gain factor is:

[0083] ;

[0084] In the formula, is the gain factor, representing the signal gain between sensor i and j at time t, and are the signal values of sensor i and j at time t respectively, is the exponential function, and are the adjustment coefficients. In this embodiment, is set to 0.5 to control the attenuation rate of the signal difference, is 0.2 to control the attenuation rate of the spatial distance.

[0085] S22. Use the graph gain method for signal fusion, construct the graph structure between the environmental electrical signals for disease prevention and optimize the signal fusion process. The mathematical model of the environmental electrical signal fusion for disease prevention is:

[0086] ;

[0087] In the formula, is the fused signal matrix optimized by graph gain, is the eigenvector matrix. By performing principal component analysis on the environmental electrical signals for disease prevention, the first 3 principal components are extracted and the energy of the principal components is retained, is the graph gain matrix. The graph gain matrix is a diagonal matrix, and the diagonal elements are the gain factors, is the relationship matrix of the environmental electrical signals for disease prevention. The relationship matrix is obtained by calculating the covariance matrix of the environmental electrical signals for disease prevention, is 's transpose matrix, is the element-wise multiplication.

[0088] S23. By combining the gain factor, graph gain, and spatial distance, and processing the signal using an exponential function, an optimized sensor signal is obtained. The mathematical model is:

[0089] ;

[0090] In the formula, is the optimized electrical signal for disease prevention environment, represents the position of sensor i, is the adjustment factor for controlling the spatial distance attenuation term. In this embodiment, is set to 0.1 to avoid overfitting, is the average value of the sensor positions.

[0091] S3. Generate multi-scale components by decomposing the signal with multi-scale fractional-order differential and wavelet basis functions, extract abnormal signal patterns, and calculate the initial abnormal probability by combining the spatial distance of the sensors and the difference in electrical signals.

[0092] Furthermore, in step S3, for calculating the initial abnormal probability, its structure is as Figure 3 shown, and the specific steps for calculating the initial abnormal probability are as follows.

[0093] S31. Decompose the electrical signal for disease prevention environment into multi-scale tensors, and directly extract potential abnormal patterns. The mathematical model is:

[0094] ;

[0095] In the formula, is the component of the electrical signal for disease prevention environment after multi-scale decomposition, is the fractional-order differential operator, , controls the decomposition of the electrical signal for disease prevention environment at different scales. s is the scale parameter, correspond to high-frequency, medium-frequency, and low-frequency components respectively, is the adaptive wavelet basis function, which controls the frequency band components through the scale parameter s;

[0096] The mathematical model of the fractional-order differential operator is:

[0097] ;

[0098] In the formula, is the gamma function, which normalizes the result of the fractional-order differential, is to take the integer part of, a is the historical time, and t is the current time point;

[0099] The mathematical model of the adaptive wavelet basis function is:

[0100] 。

[0101] S32. Combine the multi-scale components and spatial correlation to generate the initial anomaly probability. The mathematical model is:

[0102] ;

[0103] In the formula, is the initial anomaly probability, is the Euclidean space distance between sensor nodes i and j, is the smoothing constant, , to prevent the denominator from being zero, is the variance of the disease prevention environment electrical signal at the scale parameter s.

[0104] S4. Map the disease prevention environment electrical signal to the hybrid Gaussian-Riemannian manifold space, enhance the disease prevention environment electrical signal, and reverse-enhance the abnormal components through the manifold geometric gradient and fractional-order differential.

[0105] Furthermore, in step S4, for the enhanced abnormal disease prevention environment electrical signal, its structure is as Figure 4 shown. The specific steps for enhancing the disease prevention environment electrical signal are as follows.

[0106] S41. Map the disease prevention environment electrical signal to the hybrid Gaussian-Riemannian manifold to enhance the abnormal components. The mathematical model is:

[0107] ;

[0108] In the formula, is the enhanced disease prevention environment electrical signal, is the number of Gaussian components. In this embodiment, K = 3, representing three risk modes: low, medium, and high, is the uniform mixing coefficient, is the manifold center, is the covariance. The covariance and the manifold center are updated every 10 seconds through the online K-means clustering algorithm to adapt to the dynamic changes of the environment, is the adaptive fractional-order parameter, , is the information entropy, is the stability constant to prevent the denominator from being zero, .

[0109] S42. Reverse-enhance the abnormal signal through the fractional-order gradient. The mathematical model is:

[0110] ;

[0111] In the formula, is the geometric gradient of the manifold calculated by the Riemannian logarithmic mapping, , is the logarithmic mapping on the Riemannian manifold, which maps the disease prevention environmental electrical signal from the manifold space to the tangent space, is the enhancement intensity, , is when in equals the value of 0.5, is the sign function, is the assignment symbol, indicating that the signal is updated through gradient enhancement.

[0112] S5. Generate an adaptive threshold based on the mean, standard deviation, and spatial consistency metric of the local disease prevention environmental electrical signal, and combine the enhanced disease prevention environmental electrical signal with the initial anomaly probability to generate an anomaly confidence level.

[0113] Furthermore, in step S5, for calculating the anomaly confidence level, its structure is as Figure 5 shown, and the specific steps for calculating the anomaly confidence level are as follows.

[0114] S51. Generate an adaptive threshold based on the signal statistical characteristics and spatial consistency, and the mathematical model is:

[0115] ;

[0116] In the formula, is the adaptive threshold, is the local signal mean, , is the local standard deviation, , the half-width of the time window is 2, and the total length of the time window is 5. The calculation window is 5 seconds, and the mean and standard deviation of the local signal are calculated every 5 seconds, is the spatial consistency metric, is the adaptive threshold amplitude adjustment coefficient, is the spatial consistency adjustment coefficient. In this embodiment, is set to 2.0 to amplify the influence of the standard deviation on the threshold, is set to 0.5 to adjust the spatial consistency weight and avoid misjudgment of local anomalies;

[0117] The mathematical model of the spatial consistency metric is:

[0118] ;

[0119] In the formula, is the cross-correlation coefficient, which calculates the correlation of the disease prevention environmental electrical signal;

[0120] The mathematical model of the cross - correlation coefficient is as follows:

[0121] ;

[0122] In the formula, is the average value of , is the average value of , is the summation over all time instants of time t.

[0123] S52. Generate the confidence level by combining the enhanced disease prevention environment electrical signal and the adaptive threshold. The mathematical model is as follows:

[0124] ;

[0125] In the formula, is the abnormal confidence level, is the normalization constant, , and max is to take the maximum value.

[0126] S6. Fuse the confidence levels of different sensors through the Riemannian manifold distance, verify the abnormal pattern, and feedback the fusion result to the initial step S2. Dynamically adjust the fractional - order differential parameter to perform feedback optimization on the disease prevention environment electrical signal, and iteratively execute until convergence to obtain the finally processed disease prevention environment electrical signal and the final fusion confidence level.

[0127] Furthermore, in step S6, for the fusion and feedback optimization of different disease prevention environment electrical signals, its structure is as shown in Figure 6 , and the specific steps for performing feedback optimization on the disease prevention environment electrical signal are as follows.

[0128] S61. Fuse different disease prevention environment electrical signals based on the information geometric distance and verify the abnormal pattern. The mathematical model is as follows:

[0129] ;

[0130] In the formula, is the fusion confidence level after fusing the confidence levels of different sensors, is the Riemannian manifold distance, , is the fusion attenuation coefficient, , and Var is the variance.

[0131] S62. The fusion confidence level is feedback to S2 to adjust the fractional - order differential parameter . The mathematical model is as follows:

[0132] ;

[0133] In the formula, the iteration termination condition is The variance of is less than a preset threshold , and setting the preset threshold indicates that the obtained confidence level is in a stable state.

[0134] S7. Construct a disease prevention environment electrical signal processing model, input the disease prevention environment electrical signal data set, and successively go through steps S2 to S6, and complete the optimization processing of the disease prevention environment electrical signal through iterative training until convergence.

[0135] Furthermore, in step S7, for the disease prevention environment electrical signal processing model, first input the disease prevention environment electrical signal data set, and successively execute the processes in steps S2 to S6. The specific steps are as follows: fuse the disease prevention environment electrical signals by introducing a gain factor and a graph gain method; then use multi-scale fractional-order differentiation and wavelet basis function decomposition to extract abnormal signal patterns; combine manifold space mapping and fractional-order differentiation to enhance the disease prevention environment electrical signals; then generate an adaptive threshold by calculating local statistical characteristics and calculate the abnormal confidence level; then fuse the confidence levels of different sensors through the Riemannian manifold distance, verify the abnormal patterns and optimize them; the disease prevention environment electrical signal processing model is adjusted and optimized through iterative training until the processing results converge, ensuring the accuracy and stability of the optimization of the disease prevention environment electrical signals; the disease prevention environment electrical signal processing model can efficiently and accurately optimize the disease prevention environment electrical signals, improve the abnormal detection ability, and provide more reliable data support for disease prevention.

[0136] Furthermore, in step S7, for the disease prevention environment electrical signal processing model, developed based on the Python programming language, take the disease prevention environment electrical signal as the input, set the batch size to 64, use the Adam optimizer, set the learning rate to 0.001, and the number of training times to 500 times. After the iterative training is completed, the disease prevention environment electrical signal processing model outputs the final disease prevention environment electrical signal processing result.

[0137] Furthermore, in step S7, input the disease prevention environment electrical signal into the disease prevention environment electrical signal processing model for processing, and the effects before and after processing are as Figure 7As shown, the abscissa represents time in seconds, and the ordinate represents the amplitude of the signal without unit. The original disease prevention environmental electrical signal input shows obvious random noise fluctuations. On the basis of retaining the trend of the original disease prevention environmental electrical signal, the disease prevention environmental electrical signal after enhancing the abnormal components suppresses high-frequency noise and amplifies key abnormal events. The peak regions in the figure are the key abnormal events. The effects shown in the figure prove that the disease prevention environmental electrical signal processing model can effectively distinguish noise from real abnormalities, improve the signal-to-noise ratio and abnormal detection sensitivity of the disease prevention environmental electrical signal, provide reliable technical support for the processing of the disease prevention environmental electrical signal, and verify the effectiveness of the model proposed in this paper.

[0138] The above is only the preferred embodiment of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the creative concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. A disease prevention environmental sensor electrical signal processing method, characterized in that: The specific steps include: S1, collect electrical signals collected by environmental sensors and create a disease prevention environmental electrical signal dataset; S2. According to the spatiotemporal relationship of the disease prevention environment electrical signals, a gain factor is introduced, and the disease prevention environment electrical signals are fused by combining the graph gain method and the spatial distance; S3, decompose the signal by multi-scale fractional differential and wavelet basis function to generate multi-scale components, extract the abnormal signal pattern, and calculate the initial abnormal probability by combining the sensor spatial distance and electrical signal difference; S4, mapping the disease prevention environment electrical signal to the mixed Gaussian-Riemann manifold space, enhancing the disease prevention environment electrical signal, and inversely enhancing the abnormal component through the manifold geometric gradient and fractional differential; S5, generating an adaptive threshold based on the mean, standard deviation and spatial consistency measurement of the local disease prevention environment electrical signal, and generating an abnormality confidence by combining the enhanced disease prevention environment electrical signal with the initial abnormality probability; S6, fusing the confidences of different sensors through the Riemann manifold distance, verifying the abnormal mode, and feeding back the fusion results to the initial step S2, dynamically adjusting the fractional-order differential parameters, performing feedback optimization on the disease prevention environment electrical signal, iterating until convergence, and obtaining the disease prevention environment electrical signal after the final processing and the final fusion confidence; S7. Construct a disease prevention environment electrical signal processing model, input a disease prevention environment electrical signal data set, go through steps S2 to S6 in sequence, and perform iterative training until convergence to complete the optimization of the disease prevention environment electrical signal.

2. The disease prevention environmental sensor electrical signal processing method according to claim 1, characterized in that: In the S1 step, for the preparation of the disease prevention environmental electrical signal data set, environmental sensors suitable for disease prevention monitoring are selected. The environmental sensors include temperature and humidity sensors, air quality sensors, and gas concentration sensors, which are used to monitor environmental parameters such as temperature and humidity, air quality, and gas concentration. Nursing homes are selected as the collection sites for the disease prevention environmental electrical signal data sets, and sensors are installed at different monitoring locations. Environmental sensors are used to continuously collect real-time data and record electrical signals under different time and space conditions. All collected disease prevention environmental electrical signals will be stored together with the corresponding sensor sampling values ​​and spatial coordinate information, and data annotation will be performed. The disease prevention environmental electrical signals and the annotation information will be organized into a unified format to complete the preparation of the disease prevention environmental electrical signal data set.

3. The method for processing electrical signals of a disease prevention environmental sensor according to claim 2, characterized in that: In the step S2, the specific steps of fusing the disease prevention environment electrical signal are as follows: S21. Input original disease prevention environment electrical signal and sensor space coordinates , represents the sampling value of the i-th sensor node at time t, is the total number of sensor nodes, is the three-dimensional spatial coordinate of sensor i, calculate the spatial and temporal relationship between the disease prevention environment electrical signals, establish the signal association model, and obtain the gain factor between the disease prevention environment electrical signals. The mathematical model of the gain factor is: ； In the formula, is the gain factor, and are the signal values ​​of sensors i and j at time t, is an exponential function, and is the adjustment coefficient; S22. Use the graph gain method to perform signal fusion, construct the graph structure between the disease prevention environment electrical signals and optimize the signal fusion process. The mathematical model of the disease prevention environment electrical signal fusion is: ； In the formula, is the fusion signal matrix after graph gain optimization, is the eigenvector matrix, is the graph gain matrix, A relationship matrix of environmental electrical signals for disease prevention, for The transposed matrix of is element-wise multiplication; S23. By combining the gain factor, the graph gain and the spatial distance and processing the signal using an exponential function, the optimized disease prevention environmental electrical signal is obtained. The mathematical model is: ； In the formula, For optimized disease prevention environmental electrical signals, represents the position of sensor i, is the adjustment factor for controlling the spatial distance attenuation term, is the average value of the sensor positions.

4. The method for processing electrical signals of a disease prevention environment sensor according to claim 3, characterized in that: In the step S3, the specific steps of calculating the initial abnormal probability are: S31. Decompose the disease prevention environmental electrical signal into multi-scale tensors and directly extract potential abnormal patterns. The mathematical model is: ； In the formula, is the disease prevention environmental electrical signal component after multi-scale decomposition, is a fractional differential operator, Control the decomposition of disease prevention environmental electrical signals at different scales, s is the scale parameter, Corresponding to high frequency, medium frequency and low frequency components respectively, is the adaptive wavelet basis function; The mathematical model of the fractional-order differential operator is: ； In the formula, is the gamma function, To obtain The integer part of , a is the historical time, and t is the current time point; The mathematical model of adaptive wavelet basis function is: ； S32. Combine multi-scale components with spatial correlation to generate initial anomaly probability. The mathematical model is: ； In the formula, is the initial abnormal probability, is the Euclidean space distance between sensor nodes i and j, is the smoothing constant, Variance of environmental electrical signals for disease prevention under scale parameter s.

5. The disease prevention environmental sensor electrical signal processing method according to claim 4, characterized in that: In the step S4, the specific steps of enhancing the disease prevention environment electrical signal are as follows: S41. Mapping the disease prevention environmental electrical signal to a mixed Gaussian-Riemann manifold to enhance the abnormal components. The mathematical model is: ； In the formula, For enhanced disease prevention environmental electrical signals, is the number of Gaussian components, is the uniform mixing coefficient, is the center of the manifold, is the covariance, is the adaptive fractional order parameter, , is the information entropy, is the stability constant; S42, the abnormal signal is enhanced by reverse enhancement of fractional gradient, and the mathematical model is: ； In the formula, is the geometric gradient of the manifold computed via the Riemann logarithmic map, , is a logarithmic mapping on the Riemann manifold, mapping the disease prevention environmental electrical signal from the manifold space to the tangent space, To increase strength, , For In A value equal to 0.5, is the symbolic function, Is an assignment symbol, indicating that the signal is updated through gradient enhancement.

6. The disease prevention environmental sensor electrical signal processing method according to claim 5, characterized in that: In step S5, the specific steps of calculating the abnormality confidence are: S51, generating an adaptive threshold based on signal statistical characteristics and spatial consistency, the mathematical model is: ； In the formula, is the adaptive threshold, is the local signal mean, is the local standard deviation, with a calculation window of 5 seconds, is the spatial consistency measure, is the adaptive threshold amplitude adjustment coefficient, is the spatial consistency adjustment coefficient; The mathematical model of spatial consistency measurement is: ； In the formula, is the mutual correlation coefficient; The mathematical model of the mutual correlation coefficient is: ； In the formula, for The average value of for The average value of is the sum of all moments in time t; S52. The enhanced disease prevention environment electrical signal is combined with the adaptive threshold to generate abnormal confidence. The mathematical model is: ； In the formula, is the abnormal confidence level, is the normalization constant, , max is the maximum value.

7. The method for processing electrical signals of a disease prevention environment sensor according to claim 6, characterized in that: In the step S6, the specific steps of performing feedback optimization on the disease prevention environment electrical signal are as follows: S61. Based on the information geometry distance, different disease prevention environmental electrical signals are fused to verify the abnormal pattern. The mathematical model is: ； In the formula, is the fusion confidence after fusing the confidences of different sensors. is the Riemann manifold distance, , is the fusion attenuation coefficient, , Var is the variance; S62, Fusion Confidence Feedback to S2 to adjust the fractional order differential parameters , the mathematical model is: ； The iteration termination condition is The variance is less than the preset threshold.

Images