Multidirectional self-adaptive non-contact respiration monitoring method and system

Abstract

The invention relates to the technical field of intelligent Internet of Things and respiration monitoring, in particular to a multidirectional self-adaptive non-contact respiration monitoring method and system. The method comprises the following steps: firstly, collecting received data of each Rx unit in a monitoring area, preprocessing the received data, extracting CSI data of each Rx unit, and calculating a CSI ratio; then calculating a breathing energy ratio based on frequency domain transformation of CSI ratios, screening effective CSI ratios, and screening an optimal Rx unit based on counting of the effective CSI ratios; selecting an effective CSI ratio from the optimal Rx unit as a breathing waveform, and calculating an autocorrelation function of the breathing waveform; taking a period length from a starting point of the autocorrelation function to a first significant peak value position as a period length, calculating a ratio of total duration to the period length as a respiration frequency, and taking a respiration speed as a ratio of the respiration frequency to the total duration. According to the method, the defects that Wi-Fi breathing monitoring noise interference is serious and is easily influenced by the orientation of a user are overcome, and the breathing monitoring precision in the indoor environment is remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent Internet of Things and respiratory monitoring, and particularly to a multi-directional adaptive non-contact respiratory monitoring method and system. BACKGROUND

[0002] Respiration, as a basic physiological activity for maintaining life, is closely related to various physiological diseases (such as asthma, pneumonia, and pulmonary fibrosis) and psychological conditions (such as emotional stress). Continuous monitoring of key indicators such as respiratory rate, depth, and rhythm is crucial for health management, and a low-cost and efficient respiratory perception system is needed in both hospital and home environments.

[0003] In a hospital environment, respiratory monitoring is mainly aimed at critically ill and postoperative patients, which can timely detect abnormalities such as respiratory failure or pulmonary infection, helping doctors take intervention measures. In particular, patients after general anesthesia are prone to respiratory depression, and real-time monitoring can assess recovery, reduce the risk of complications, and alleviate the burden on medical staff, improving the efficiency of medical resources. In a home environment, respiratory monitoring is also indispensable. Patients with sleep apnea syndrome often experience apnea at night, leading to a decrease in blood oxygen and increasing the risk of cardiovascular disease, and continuous monitoring can help early detection and intervention. Chronic respiratory disease patients such as chronic obstructive pulmonary disease and pulmonary fibrosis can better understand their health status through long-term monitoring, and doctors can optimize treatment plans accordingly. With the intensification of population aging in China, the respiratory health problems of the elderly living alone are becoming increasingly prominent. The elderly are prone to chronic bronchitis, pneumonia, and other diseases, and are more likely to experience acute exacerbations in winter. An intelligent respiratory monitoring system can track respiratory status in real time and provide early warnings, helping family members or community medical personnel intervene in a timely manner, effectively reducing the risk of sudden health events.

[0004] Therefore, whether in a hospital or a home environment, a convenient and efficient respiratory monitoring system has important value. It not only can improve medical quality and optimize chronic disease management, but also can provide health protection for the elderly. With the development of sensing technology and artificial intelligence, future respiratory monitoring systems will be more intelligent, bringing health benefits to a wider population.

[0005] Respiratory monitoring technology is mainly divided into two categories: contact and non-contact. Contact monitoring is further divided into invasive and non-invasive, with the former being a cannula-type airflow sensor and the latter being a chest strap-type respiratory sensor. Although these methods have high accuracy, they generally have problems such as discomfort, limited user activity, and high cost, making it difficult to meet the unobtrusive monitoring needs in daily environments. Non-contact monitoring is achieved through acoustics, optics, or wireless signals, avoiding direct contact with users, making it more suitable for long-term and continuous respiratory monitoring.

[0006] In non-contact monitoring, methods based on traditional physical information (such as acoustics, optics) rely on expensive equipment such as cameras or lidar, and are easily disturbed by environmental light or noise. Privacy issues also limit their application. In contrast, wireless sensing-based technologies (such as radar, RFID, Wi-Fi) have become a research hotspot due to their non-contact, low-cost, and easy-to-deploy characteristics. Among them, Wi-Fi technology has unique advantages due to its widespread popularity and signal penetration capabilities. Wi-Fi-based respiratory monitoring technology has become a research hotspot due to its unique advantages. First, Wi-Fi signals are widely available in everyday environments, eliminating the need for additional deployment of dedicated equipment, significantly reducing system costs. Second, Wi-Fi signals have good penetration, covering multiple rooms indoors and are suitable for large-scale monitoring. In addition, Wi-Fi monitoring is completely non-contact and does not interfere or discomfort users, making it ideal for long-term, continuous home respiratory monitoring. Compared with other non-contact technologies (such as radar or optical methods), Wi-Fi solutions have advantages in cost, deployment convenience, and privacy protection, making them an ideal choice for smart home and health monitoring scenarios.

[0007] Currently, Wi-Fi respiratory monitoring technology still faces some key problems. Hardware noise and signal attenuation are particularly prominent, and the weak signal changes caused by breathing are easily masked by device noise, especially in long distances or complex environments. In addition, Wi-Fi signals are highly sensitive to the user's position and orientation, and the monitoring results differ greatly under different postures or non-line-of-sight paths. SUMMARY

[0008] In order to overcome the defects of Wi-Fi respiratory monitoring noise interference and being easily affected by user orientation in the prior art, the present application provides a multi-directional adaptive non-contact respiratory monitoring method, which significantly improves the respiratory monitoring accuracy in indoor environments.

[0009] The multi-directional adaptive non-contact respiratory monitoring method provided by the present application comprises the following steps: S1, collecting and preprocessing the received data of each Rx unit in the monitoring area, extracting the CSI data of each Rx unit, and calculating the CSI ratio; S2, calculating the respiratory energy proportion based on the frequency domain transformation of the CSI ratio, screening the effective CSI ratio, and selecting the best Rx unit based on the effective CSI ratio count; S3, selecting an effective CSI ratio from the best Rx unit as the respiratory waveform, and calculating its autocorrelation function; S4, taking the start point of the autocorrelation function to the position of the first significant peak as the period length, and calculating the ratio of the total duration to the period length as the respiratory rate. The respiratory rate is the ratio of the respiratory rate to the total duration.

[0010] Preferably, the CSI ratio respiratory energy proportion in step S2 is calculated in the following manner: first, frequency transform is performed on the CSI ratio to obtain a frequency curve, and then the respiratory energy proportion of the CSI ratio is calculated according to the following formula: ; Wherein, V(f) is the amplitude corresponding to the frequency f on the frequency curve.

[0011] Preferably, in step S2, the CSI ratio corresponding to the respiratory energy proportion greater than or equal to the set threshold is obtained as the effective CSI ratio, and the Rx unit with the largest number of effective CSI ratios is selected as the best Rx unit.

[0012] Preferably, in step S3, the effective CSI ratio with the largest variance is selected as the respiratory waveform.

[0013] Preferably, in step S1, the pre-processing of the received data is performed in the following manner: the CSI data of each Rx unit is extracted and time sequence alignment is performed; and then the outliers of each CSI data are removed.

[0014] Preferably, in step S1, after the pre-processing of the CSI data, the CSI ratio is calculated and the CSI ratio exceeding the set CSI ratio interval is deleted.

[0015] Preferably, in step S2, fast Fourier transform is used for frequency domain transformation.

[0016] The system of the multi-directional adaptive non-contact respiratory monitoring method proposed by the application comprises a data pre-processing module, a subcarrier and an Rx screening module, and a respiratory rate calculation module. The data pre-processing module is used for collecting and pre-processing the received data of each Rx unit in the monitoring area, and calculating the CSI ratio. The subcarrier and the Rx screening module are used for screening the effective CSI ratio of each Rx unit, and screening the best Rx unit based on the effective CSI ratio count. The respiratory rate calculation module screens the respiratory waveform from the effective CSI ratio of the best Rx unit, and calculates the respiratory rate.

[0017] The multi-directional adaptive non-contact respiratory monitoring device proposed by the application comprises a memory and a processor, the memory stores a computer program, the processor is connected to the memory, and the processor is used to execute the computer program to realize the multi-directional adaptive non-contact respiratory monitoring method.

[0018] The storage medium proposed by the application stores a computer program, and the computer program is executed to realize the multi-directional adaptive non-contact respiratory monitoring method.

[0019] The advantages of the present application are: (1) In the multi-directional self-adaptive non-contact respiratory monitoring method proposed by the present application, the CSI ratio screening scheme based on the respiratory energy threshold can effectively identify the CSI ratio highly correlated with the respiratory signal, accurately calculate the user's respiratory rate, and significantly improve the signal quality. The receiving unit screening scheme based on counting analysis quantitatively evaluates the sensing ability of each node, realizes intelligent selection of the best receiving unit, and thus can adaptively sense the respiratory signal of the user in different positions and orientations in an indoor environment, ensuring that the respiratory signal can be accurately captured even in complex environmental conditions, greatly enhancing the stability of monitoring in different positions and orientations, and adapting to various indoor scenes. Experiments show that the present application can maintain stable monitoring performance when the user is in different positions and orientations, showing excellent adaptability and robustness. (2) The CSI ratio dynamic screening mechanism based on the respiratory power ratio threshold constructed in the present application quantitatively analyzes the differences in the respiratory signal representation ability of different CSI ratios, establishes an adaptive CSI ratio optimization criterion, and fundamentally solves the technical problem of uneven signal quality of multiple sub-carrier signals, greatly improving the signal-to-noise ratio of the original signal and making a breakthrough in the accuracy of extracting respiratory waveform features.

[0020] (3) The receiving unit quantitative evaluation system based on the number of CSI ratio screening constructed in the present application establishes a mapping model between the effective utilization rate of CSI ratio and the node sensing sensitivity, and for the first time realizes the standardized measurement of the performance of the receiving unit, providing scientific basis and universal guiding principles for key decisions such as node optimization and spatial layout optimization during system deployment, significantly improving the system robustness of multi-node collaborative sensing in complex scenarios, and solving the practical challenge of significant differences in the sensing performance of each node in a multi-receiving system.

[0021] (4) The present application can use small-volume and low-cost commercial Wi-Fi devices such as ESP32 to construct an extensible one-transmitting and multiple-receiving sensing architecture, which has better popularity and scalability, and provides a technical path for large-scale deployment and landing application of Wi-Fi respiratory sensing system in actual scenarios. Through the combination of low-cost commercial Wi-Fi devices and the CSI ratio dynamic screening mechanism based on the respiratory power ratio threshold, the inherent problems of large signal noise and poor stability of Wi-Fi hardware platform are overcome, realizing a low-cost, low-noise, and high-performance sensing system.

[0022] (5) This invention has significant advantages in terms of cost control, performance, and environmental adaptability. Through this invention, low-cost commercial equipment can be used, making it potentially deployable on a large scale. The algorithm design of this invention ensures the accuracy and stability of monitoring. These advantages make this system particularly suitable for applications in scenarios such as ward monitoring and home health management, providing a new solution for promoting the popularization of respiratory monitoring technology.

[0023] (6) The multi-directional adaptive non-contact respiratory monitoring system proposed in this invention is essentially a non-contact respiratory monitoring system based on the Single Transmitter, Multiple Receivers (SCR) mode. It achieves comprehensive coverage of the Wi-Fi respiratory sensing system throughout the entire indoor space and ensures its suitability for users in any orientation and location. While ensuring sensing accuracy, the system also considers the cost and flexibility of device deployment. This invention selects the compact, inexpensive ESP32 development board with CSI acquisition capabilities as the core sensing device, reducing deployment costs at the hardware level and making it possible to build a multi-node, fully covered indoor respiratory sensing network. Attached Figure Description

[0024] Figure 1 This is a flowchart of a multi-directional adaptive non-contact respiratory monitoring method proposed in this invention; Figure 2 A block diagram of a multi-directional adaptive non-contact respiratory monitoring system; Figure 3(a) shows the subcarriers; Figure 3(b) shows another subcarrier; Figure 3(c) shows the CSI ratio; Figure 3(d) shows the frequency curve obtained from Figure 3(c) after FFT transformation; Figure 4 For all CSI ratios on the selected optimal Rx; Figure 5(a) shows the respiratory waveform; Figure 5(b) shows the combined autocorrelation coefficients of the respiratory waveform; Figure 6 This is a schematic diagram of the experimental scenario; Figure 7 This represents the absolute error for different orientations at different locations; Figure 8 This represents the cumulative distribution of absolute errors; Figure 9 The detection rate exists for respiration at different locations. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0026] like Figure 1 As shown in the figure, the multi-directional adaptive non-contact respiratory monitoring method proposed in this embodiment includes the following steps: S1. Collect and preprocess the received data of each Rx unit (WIFI receiving unit) within the monitoring area, extract the CSI (channel state information) data of each Rx unit, and calculate the CSI ratio.

[0027] The preprocessing method includes the following steps: First, extract the CSI data (i.e., the CSI amplitude curve) and time-align the CSI data of each Rx unit; then remove outliers from each CSI data and perform filtering; finally, calculate the CSI ratio within each Rx unit.

[0028] Each Rx unit receives data containing multiple CSI data points, and each CSI data point contains multiple subcarriers, as shown in Figures 3(a) and 3(b). During preprocessing, an outlier removal method based on statistical characteristics is used to remove outliers from each CSI data point. Then, hardware noise is removed based on the subcarrier ratio. Specifically, for each CSI data point received by the Rx unit, the ratio between the subcarriers is calculated as the CSI ratio, as shown in Figure 3(c). A filter is then applied based on a set CSI ratio range to remove CSI ratios that exceed this range, thus achieving CSI ratio filtering.

[0029] S2. Calculate the respiratory energy ratio based on the frequency domain transformation of the CSI ratio, screen the effective CSI ratio, and screen the best Rx unit based on the count of the effective CSI ratio.

[0030] This step S2 specifically includes the following sub-steps S21-S24.

[0031] S21. Perform frequency transformation on the CSI ratio of the Rx unit. Specifically, a fast Fourier transform can be used to obtain the frequency curve shown in Figure 3(d). S22. Calculate the proportion of respiratory energy on the frequency curve. The calculation formula is as follows: ; V(f) is the amplitude corresponding to frequency f on the frequency curve.

[0032] The respiratory frequency band is set according to the actual situation. In the embodiment shown in Figure 3(d), the respiratory rate is set to a range of 10 breaths / minute to 30 breaths / minute, that is, the respiratory frequency band is [10, 30]. .

[0033] S23. Obtain the CSI ratios that are greater than or equal to a set threshold as effective CSI ratios. Sum the number of effective CSI ratios for each CSI data in the Rx unit to obtain the number of effective CSIs in the Rx unit. S24. Select the Rx cell with the most effective CSI ratios as the optimal Rx cell. The effective CSI ratio of a certain optimal Rx cell is as follows: Figure 4 As shown.

[0034] S3. Select an effective CSI ratio from the best Rx unit as the breathing waveform, smooth it, and then calculate the autocorrelation function, as shown in Figure 5(a) and Figure 5(b).

[0035] In practice, the effective CSI ratio with the largest variance can be selected as the respiratory waveform in this step.

[0036] S4. Using the period from the starting point of the autocorrelation function to the position of the first significant peak as the period length, calculate the ratio of the total duration to the period length as the total number of periods, i.e., the number of breaths; calculate the ratio of the number of breaths to the total duration as the respiratory rate.

[0037] In this step, significant peaks can be directly detected using AMPD (peak detection algorithm).

[0038] Reference Figure 2 This embodiment proposes a multi-directional adaptive non-contact respiratory monitoring system, which includes: a data preprocessing module, a subcarrier and Rx screening module, and a respiratory rate calculation module.

[0039] The data preprocessing module is used to collect and preprocess the received data from each Rx unit (WIFI receiving unit) within the monitoring area and calculate the CSI ratio.

[0040] The subcarrier and Rx filtering module is used to filter the effective CSI ratio of each Rx unit and filter the best Rx unit based on the effective CSI ratio count.

[0041] The respiratory rate calculation module filters respiratory waveforms from the effective CSI ratio of the optimal Rx unit and calculates the respiratory rate.

[0042] The above-described multi-directional adaptive non-contact respiratory monitoring method is verified in conjunction with specific embodiments below.

[0043] This embodiment deploys the method in the real world to evaluate performance. In the experimental environment, six Wi-Fi receiver units Rx are arranged around the Wi-Fi transmitter unit Tx. For example... Figure 6 As shown, to verify the respiratory perception performance of the system in an indoor environment, six spatially representative monitoring points, Loc1 to Loc6, were selected in the experimental setting. At each monitoring point, subjects performed a seated test facing four different directions (up, down, left, and right). Ten sets of data were collected for each direction, resulting in a total of 240 sets of multidimensional CSI data sequences. Furthermore, to ensure data validity, the arrangement of personnel and equipment in the experimental setting remained consistent in each experiment, ensuring repeatability and data reliability.

[0044] In this experiment, when calculating the proportion of respiratory power for different CSI ratios, the following steps are taken: First, a Fast Fourier Transform (FFT) is performed on the CSI ratio using the FFT function to calculate the amplitude at different frequencies. Next, the window size is determined based on the product of the sampling rate and the sampling time, providing a basis for the frequency range. Then, based on the FFT results, the amplitude and frequency range corresponding to each frequency point are calculated. The algorithm iterates through all possible frequency points and their corresponding amplitudes, obtaining the total power by accumulating the squares of the amplitudes. Simultaneously, the algorithm determines whether the current frequency is within the respiratory rate range (typically 10 to 30 breaths / minute); if so, the power at that frequency point is added to the respiratory power. Finally, the algorithm calculates the proportion of respiratory power by dividing the respiratory power by the total power. If this proportion is greater than a preset threshold, the corresponding CSI ratio is added to the filtering results.

[0045] This experiment employs a method combining autocorrelation coefficient (ACR) and the AMPD algorithm to calculate respiratory rate. This method leverages the advantages of both algorithms to achieve high-precision identification and calculation of the respiratory cycle. ACR effectively captures the periodic characteristics of the respiratory signal, while the AMPD algorithm accurately detects peak points in the signal through multi-scale analysis, thus avoiding misjudgments that might occur with a single algorithm. This dual detection mechanism significantly enhances the system's anti-interference capability, effectively suppressing spurious peak interference caused by environmental noise or signal fluctuations, while also optimizing the accuracy of cycle length determination. Experimental results show that the introduction of this method not only significantly improves the accuracy of respiratory rate calculation but also enhances the system's adaptability in complex scenarios, providing more reliable technical support for respiratory monitoring in practical applications.

[0046] The results of this experiment are as follows Figures 7-9 As shown. From Figure 7As can be seen, the average error at all test locations was controlled within 0.5 breaths / minute, with the largest average error at the location being 0.48 breaths / minute. Overall, the system's average error performance was particularly excellent, reaching 0.31 breaths / minute, exceeding the requirements of ISO 80601-2:2017 for the performance of normal medical devices. This result demonstrates that the system maintains high-precision respiratory sensing capabilities under different positions and orientations, making it feasible for practical applications.

[0047] from Figure 8 As can be seen, the cumulative distribution function of the respiratory rate measurement error of the present invention shows that the error of 91.3% of the measurement samples is controlled within the clinical threshold of 0.5 breaths / minute. In particular, the error of approximately 98% of the samples is less than 1 breath / minute, further demonstrating the stability and high accuracy of the system under various experimental conditions.

[0048] from Figure 9 As can be seen, except for Loc5, the detection rate of the other five positions reached 100% in all orientations. The only case of detection failure was when the subject was facing upwards (towards the wall) at the Loc5 position. This position is extremely close to the wall and the direction is towards the wall, which is a relatively extreme signal occlusion condition. This may lead to a severe weakening of the signal multipath effect, thus affecting the judgment of the screening algorithm.

[0049] Nevertheless, the overall average detection rate of the system at all test points still reached over 96%, which fully demonstrates that the system has a high degree of adaptability and stable respiratory detection capability in normal indoor environments, and can meet the needs of daily respiratory monitoring and health sensing applications.

[0050] Of course, those skilled in the art will recognize that the present invention is not limited to the details of the exemplary embodiments described above, but also includes the same or similar structures that can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0051] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0052] The technologies, shapes, and structures not described in detail in this invention are all known technologies.

Claims

1. A multi-directional adaptive non-contact respiratory monitoring method, characterized in that, Includes the following steps: S1. Collect and preprocess the received data of each Rx unit within the monitoring area, extract the CSI data of each Rx unit, and calculate the CSI ratio. S2. Calculate the respiratory energy ratio based on the frequency domain transformation of the CSI ratio, screen the effective CSI ratio, and screen the best Rx unit based on the count of the effective CSI ratio; S3. Select an effective CSI ratio from the best Rx unit as the breathing waveform and calculate its autocorrelation function; S4. Using the period from the starting point of the autocorrelation function to the position of the first significant peak as the period length, calculate the ratio of the total duration to the period length as the number of breaths, and the breathing rate as the ratio of the number of breaths to the total duration.

2. The multi-directional adaptive non-contact respiratory monitoring method as described in claim 1, characterized in that, The calculation method for the respiratory energy percentage of the CSI ratio in step S2 is as follows: First, the CSI ratio is frequency-transformed to obtain a frequency curve, and then the respiratory energy percentage of the CSI ratio is calculated according to the following formula: Where V(f) is the amplitude corresponding to frequency f on the frequency curve.

3. The multi-directional adaptive non-contact respiratory monitoring method as described in claim 2, characterized in that, In step S2, the CSI ratio with the corresponding respiratory energy percentage greater than or equal to the set threshold is obtained as the effective CSI ratio, and the Rx unit with the most effective CSI ratios is selected as the best Rx unit.

4. The multi-directional adaptive non-contact respiratory monitoring method as described in claim 1, characterized in that, In step S3, the effective CSI ratio with the largest variance is selected as the respiratory waveform.

5. The multi-directional adaptive non-contact respiratory monitoring method as described in claim 1, characterized in that, The method for preprocessing the received data in step S1 is as follows: extract the CSI data of each Rx unit and perform timing alignment; then remove outliers from each CSI data.

6. The multi-directional adaptive non-contact respiratory monitoring method as described in claim 5, characterized in that, In step S1, after preprocessing the CSI data, the CSI ratio is calculated and CSI ratios that exceed the set CSI ratio range are deleted.

7. The multi-directional adaptive non-contact respiratory monitoring method as described in claim 1, characterized in that, In step S2, a fast Fourier transform is used to perform frequency domain transformation.

8. A system carrying out the multi-directional adaptive non-contact respiratory monitoring method as described in any one of claims 1-7, characterized in that, include: Data preprocessing module, subcarrier and Rx filtering module, and respiratory rate calculation module; The data preprocessing module is used to collect the received data from each Rx unit within the monitoring area, preprocess it, and calculate the CSI ratio. The subcarrier and Rx filtering module is used to filter the effective CSI ratio of each Rx unit and filter the best Rx unit based on the effective CSI ratio count; The respiratory rate calculation module filters respiratory waveforms from the effective CSI ratio of the optimal Rx unit and calculates the respiratory rate.

9. A multi-directional adaptive non-contact respiratory monitoring device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, the processor is connected to the memory, and the processor is used to execute the computer program to implement the multi-directional adaptive non-contact respiratory monitoring method as described in any one of claims 1-7.

10. A storage medium, characterized in that, The device contains a computer program that, when executed, is used to implement the multi-directional adaptive non-contact respiratory monitoring method as described in any one of claims 1-7.

Images